2013-06-01 14:55:17 hi 2013-06-01 15:03:03 How can I test I/O on alpine? 2013-06-01 19:39:43 hdparm I guess 2013-06-01 19:39:45 or just dd 2013-06-01 19:40:29 ah, he quitted 2013-06-01 19:41:37 which file will be executed every bootup? 2013-06-01 19:41:51 or not 2013-06-01 19:41:53 need to create a directory every bootup in /var/run 2013-06-01 19:41:56 tab-completion failed… 2013-06-01 19:42:30 I guess there is kind of rc.local, but probably named differently 2013-06-01 19:43:38 let me take a look 2013-06-01 19:44:04 what about /etc/init.d/local and /etc/local.d? 2013-06-01 19:44:12 not sure if we support tmpfiles.d 2013-06-01 19:44:14 ncopa: ^ 2013-06-01 19:44:29 We do 2013-06-01 19:44:37 barthalion: i will try, thanks 2013-06-01 19:44:41 read then about tmpfiles.d support in openrc 2013-06-01 19:44:44 it's better method 2013-06-01 19:44:53 I mean correct one, both will work 2013-06-01 19:48:30 we have /etc/local.d 2013-06-01 19:52:41 yeah, but it's not its purpose 2013-06-01 19:53:55 just put *.conf in /etc/tmpfiles.d, it should work 2013-06-01 19:54:05 I don't understand why we don't use it 2013-06-01 19:55:36 http://sprunge.us/fKPO this is how could look /etc/tmpfiles.d/mpd.conf 2013-06-01 19:55:54 obviously it creates /run/mpd dir with 0755, and mpd:mpd is the owner 2013-06-01 19:57:31 hm, you may need to add tmpfiles.setup to some runlevel 2013-06-01 19:58:52 we should really discuss making it default someday 2013-06-01 19:59:05 instead of using mkdir in daemons… 2013-06-01 20:04:01 shafire: additionally local.d should be executed after all daemons 2013-06-01 20:04:23 if not, I drink your health, openrc devs 2013-06-01 20:06:42 i have a bigger problem :-( 2013-06-01 20:06:53 I get always a segmentation fault, when I start a xen guest 2013-06-01 20:07:04 and if I poweroff a xen guest 2013-06-01 20:07:15 but everything works well :S 2013-06-01 20:08:03 I don't use Xen under Alpine, you should wait for someone who do 2013-06-01 20:08:13 royger or kaniini probably 2013-06-01 20:09:11 Do you use XEN? 2013-06-01 20:09:27 Yeah, but I ran Debian Wheezy on my dom0 2013-06-01 20:09:35 not my choice unfortunately 2013-06-01 20:09:40 and your domU are alpine? 2013-06-01 20:09:54 nope 2013-06-01 20:10:31 if I have some spare memory to assign I can make a new domU 2013-06-01 20:10:57 hrm, I don't, sorry 2013-06-01 20:11:56 [ 622.083418] grsec: Segmentation fault occurred at 0000000000000001 in /usr/sbin/xl[xl:2076] uid/euid:0/0 gid/egid:0/0, parent /bin/busybox[ash:1531] uid/euid:0/0 gid/egid:0/0 2013-06-01 20:12:05 shafire: ivy bridge? 2013-06-01 20:12:11 kaniini: yes 2013-06-01 20:12:38 is my ivy bridge a problem? 2013-06-01 20:12:38 shafire: i'm not 100% sure why that happens 2013-06-01 20:13:16 shafire: no, it works in practice 2013-06-01 20:13:40 http://pastie.org/private/izuxtrklv7nyxdibnwwiw 2013-06-01 20:13:47 i have a ton of ivy bridges running alpine 2013-06-01 20:13:50 i get segfaults in xl 2013-06-01 20:13:52 but it works anyway 2013-06-01 20:14:09 yes 2013-06-01 20:14:19 do you use xl? 2013-06-01 20:14:22 yes 2013-06-01 20:14:28 or switched to xm because of segfaults? 2013-06-01 20:14:28 are you using udev? 2013-06-01 20:14:34 no, i am using xl 2013-06-01 20:14:44 I don't know, if I am using it... How can I check? 2013-06-01 20:14:51 ps aux | grep udev 2013-06-01 20:15:10 2156 root 0:00 grep udev 2013-06-01 20:15:13 what's udev? 2013-06-01 20:15:19 interesting 2013-06-01 20:15:25 you say your guests are working? 2013-06-01 20:15:29 yes 2013-06-01 20:15:42 grep run_hotplug_scripts /etc/xen/xl.conf 2013-06-01 20:15:44 I get segfaults on startup and poweroff 2013-06-01 20:15:49 yeah 2013-06-01 20:15:51 i do too 2013-06-01 20:15:53 it's a known issue 2013-06-01 20:16:01 #run_hotplug_scripts=1 2013-06-01 20:16:06 okay 2013-06-01 20:22:36 Is it because of ivy bridge or alpine 2.6.0? 2013-06-01 20:25:17 kaniini: do you use lvm for your xen guest hard disks? 2013-06-01 20:25:26 and same question also to barthalion 2013-06-01 20:25:38 no 2013-06-01 20:26:07 just files? 2013-06-01 20:26:29 yeah, it probably wasn't the best idea 2013-06-01 20:27:14 but easiest, I didn't have enough knowledge then 2013-06-01 20:27:34 I can not get flashcache run with LVM :-( Or the flashcache stats are broken... 2013-06-01 20:30:55 no, it works 2013-06-01 20:30:56 :D 2013-06-01 20:31:04 :p 2013-06-01 20:31:08 first time: Timing buffered disk reads: 334 MB in 3.01 seconds = 111.08 MB/sec 2013-06-01 20:31:17 5 rounds later: Timing buffered disk reads: 698 MB in 3.01 seconds = 231.79 MB/sec 2013-06-01 20:31:36 75% cache read hit :-) 2013-06-01 20:32:48 Okay, very good. 2013-06-01 20:32:59 About 90% cache hit rate and 240MB/s :-) 2013-06-01 20:33:18 Now, how can I test I/O ... mh 2013-06-01 20:35:59 xl jumps to 0x1 2013-06-01 20:37:12 which error message is 0x1? 2013-06-01 20:37:25 which means... 2013-06-01 20:37:28 NULL deref. 2013-06-01 20:37:33 i am guessing... 2013-06-01 20:37:41 unsigned char cpu_id; 2013-06-01 20:37:48 void *some_function_pointer; 2013-06-01 20:37:57 some_function_pointer is 0x00000000001 2013-06-01 20:39:45 [ 585.282055] xl[1792]: segfault at 1 ip 00007ad9054faacb sp 00007ad9078a23c0 error 4 in libgcc_s.so.1[7ad9054ef000+11000] 2013-06-01 20:39:53 something to do with libgcc_s.so? 2013-06-01 20:42:55 or this: [ 608.116385] grsec: denied resource overstep by requesting 4096 for RLIMIT_CORE against limit 0 for /usr/sbin/xl[xl:1935] uid/euid:0/0 gid/egid:0/0, parent /bin/busybox[ash:1531] uid/euid:0/0 gid/egid:0/0 2013-06-01 20:45:44 no 2013-06-01 20:46:11 i'll figure it out monday 2013-06-01 20:46:22 need to talk to royger 2013-06-01 20:46:51 okay 2013-06-01 20:50:10 Should I open a bug for this, or could you do it please, kaniini? 2013-06-01 20:52:36 i will just talk to royger about it monday. 2013-06-01 20:54:06 ok 2013-06-01 21:13:11 xc: debug: hypercall buffer: cache hits:36 misses:2 toobig:1 2013-06-01 21:13:13 Segmentation fault 2013-06-01 21:13:17 hm 2013-06-01 21:13:54 Was it you with the upgrade to 4.2.2? 2013-06-01 21:14:03 algitbot: retry build64-edge 2013-06-01 21:19:59 shafire: yes 2013-06-01 21:24:13 kaniini: could you please check my bug with the /var/run/xen 2013-06-01 21:30:05 openrc has tmpfiles.d, but we don't 2013-06-01 21:30:34 it requires some manual steps. 2013-06-01 21:30:50 shafire: that is also a known issue 2013-06-01 21:32:58 crash happens after xc__hypercall_buffer_cache_release(). 2013-06-01 21:35:07 does logrotate really need to be a depend for dovecot 2013-06-01 21:41:18 pod2text man/xl.pod.1 txt/man/xl.1.txt.tmp 2013-06-01 21:41:20 man/xl.pod.1 around line 854: Expected text after =item, not a bullet 2013-06-01 21:41:22 POD document had syntax errors at /usr/bin/pod2text line 84. 2013-06-01 21:41:24 are you fucking kidding me 2013-06-01 22:04:28 okay 2013-06-01 22:04:33 shafire: i think i figured it out 2013-06-01 22:20:16 kaniini: juhu! 2013-06-01 22:27:00 problem is not in xen, problem is in uclibc. 2013-06-01 22:27:08 i have a testcase 2013-06-01 22:27:33 will work on fix as soon as i get xen 4.2.2 build fixed 2013-06-01 22:35:13 kaniini: are you using hyperthreading in your ivy bridges? 2013-06-01 22:37:20 yes 2013-06-01 22:37:28 i have a bunch of e3-1270v2's 2013-06-01 22:38:02 I have obe e3-1265v2 2013-06-01 22:38:05 one* 2013-06-01 22:38:47 i have about 60. 2013-06-01 22:38:49 :P 2013-06-01 22:41:57 are you a sysadmin? 2013-06-01 22:43:26 the build log is 5.9mb big and 99% is rubbish? 2013-06-01 22:44:35 that should get that done 2013-06-01 22:45:50 shafire: yes/no 2013-06-01 22:49:07 should I try it? 2013-06-01 22:51:15 kaniini: or what do you mean with yes/no? 2013-06-01 22:51:40 now, I can view the log :) 2013-06-01 23:12:06 (1/1) Upgrading libpthread (0.9.33.2-r21 -> 0.9.33.2-r22) 2013-06-01 23:12:09 lets see if this works 2013-06-01 23:12:27 random linux question 2013-06-01 23:12:39 localhost:/home/kaniini# xl destroy lol 2013-06-01 23:12:41 localhost:/home/kaniini# xl create /etc/xen/nenolod-10.cfg 2013-06-01 23:12:43 Parsing config from /etc/xen/nenolod-10.cfg 2013-06-01 23:12:43 can only stuff running as root write their pidfiles to /var/run ? 2013-06-01 23:12:45 Daemon running with PID 13539 2013-06-01 23:12:48 no more segfault ! 2013-06-01 23:17:25 Kabaka: this probably fixes your ruby segfault problem too 2013-06-01 23:17:29 Kabaka: ^^^^ 2013-06-01 23:17:54 :D 2013-06-01 23:18:16 :D 2013-06-01 23:18:24 Kabaka: ruby segfault? 2013-06-01 23:18:29 when you got them? 2013-06-01 23:20:04 shafire: do this 2013-06-01 23:20:18 shafire: echo '@edge http://nl.alpinelinux.org/alpine/edge/main' >> /etc/apk/repositories 2013-06-01 23:20:22 shafire: apk update 2013-06-01 23:20:26 shafire: apk add libpthread@edge 2013-06-01 23:20:30 shafire: apk fix libpthread 2013-06-01 23:20:36 shafire: your segfaults in xl should go away 2013-06-01 23:20:46 ok 2013-06-01 23:20:50 I will try 2013-06-01 23:21:49 what is @edge? 2013-06-01 23:22:18 what i just pushed that libc change too 2013-06-01 23:22:51 jippii 2013-06-01 23:22:51 localhost:~# xl create /srv/vm0.pv Parsing config from /srv/vm0.pv Daemon running with PID 2509 localhost:~# 2013-06-01 23:22:54 it's gone 2013-06-01 23:22:55 :-) 2013-06-01 23:23:21 okay 2013-06-01 23:23:26 i'll backport it to 2.6 2013-06-01 23:24:00 What was wrong? 2013-06-01 23:24:04 I am really interested 2013-06-01 23:28:54 shafire: unwind functions track cpu registers state 2013-06-01 23:29:01 shafire: if unwind is too old, ... 2013-06-01 23:29:28 i would rather have the slightly slower version which tracks cleanup callbacks in a linked list 2013-06-01 23:29:31 and is safe 2013-06-01 23:29:53 ah okay 2013-06-01 23:30:24 Kabaka: what did you do with ruby that you got segmentation faults? 2013-06-01 23:30:36 shafire: ruby uses threads internally heavily. 2013-06-01 23:31:00 the reason xen reliably triggered the crash btw 2013-06-01 23:31:03 pthread_atexit() 2013-06-01 23:31:45 shafire: a Sinatra application was segfaulting at completely random places, often in string manipulation functions 2013-06-01 23:32:37 yeah :P 2013-06-01 23:32:55 Kabaka: same thing here -- completely junk backtraces 2013-06-01 23:33:01 kaniini: I got in all test scenoaries zero segmentation faults :-) 2013-06-01 23:33:44 pthread_cleanup_* functions also reliably trigger it 2013-06-01 23:34:58 So, you saved me time to test KVM :P 2013-06-01 23:35:16 Will stay with XEN now 2013-06-01 23:36:10 but I think, I would have also segmentation faults with KVM 2013-06-01 23:40:43 kaniini: If I do now apk upgrade, will I get all edge packages? 2013-06-01 23:40:58 Should I remove the @edge line 2013-06-01 23:40:58 ? 2013-06-01 23:41:44 no 2013-06-01 23:41:55 the @edge part means it's a tagged repository 2013-06-01 23:42:08 which means you can only pin things in it 2013-06-01 23:42:21 however... 2013-06-01 23:42:26 you should remove it 2013-06-01 23:42:28 then... 2013-06-01 23:42:33 apk add libpthread 2013-06-01 23:42:36 apk fix libpthread 2013-06-01 23:42:38 :) 2013-06-01 23:42:44 then you'll get the 2.6 backport 2013-06-01 23:43:31 I am using the hetzner server, I don't know, if the server is already synced 2013-06-01 23:43:52 apk update will sync it 2013-06-01 23:43:56 hmm 2013-06-01 23:43:58 oic 2013-06-01 23:43:59 well 2013-06-01 23:44:03 wait a day or two 2013-06-01 23:44:05 :P 2013-06-01 23:44:08 ;-) 2013-06-01 23:44:10 (hetzner server ?) 2013-06-01 23:44:16 yeah 2013-06-01 23:44:32 dl-5.alpinelinux.org is a hetzner server 2013-06-01 23:45:39 oh. okay. 2013-06-01 23:45:40 Is xen 4.2.2 now available too? 2013-06-01 23:45:46 on edge, it is. 2013-06-01 23:45:56 that will not be backported to 2.6. 2013-06-01 23:46:31 okay 2013-06-01 23:46:50 edge means, it will be available in the next version (2.7?) 2013-06-01 23:48:02 yes. 2013-06-01 23:48:08 but, xen 4.3 will be in 2.7 probably :) 2013-06-01 23:49:22 mh, okay. but why will be 4.2.2 not be backported? it's only a minor upgrade, or am I wrong? 2013-06-01 23:49:47 http://www.xenproject.org/downloads/xen-archives/supported-xen-42-series/xen-422.html 2013-06-01 23:50:02 We recommend that all users of Xen 4.2.1 upgrade to Xen 4.2.2. -> :-( 2013-06-01 23:50:59 all of the security fixes are backported. 2013-06-01 23:51:17 stable releases do not get new versions unless there is some specific reason to do so 2013-06-01 23:52:03 ah ok 2013-06-01 23:52:22 thank you 2013-06-01 23:52:41 if you'd like 4.2.2 to be backported to 2.6, request it on alpine-devel and if ncopa says it is okay, i will do it :P 2013-06-01 23:53:08 i only did the libc backport because there were many realworld apps breaking in weird ways on ivy bridge cpus 2013-06-01 23:53:15 so, it had to be done ;) 2013-06-01 23:53:20 yeah 2013-06-01 23:53:56 have fun updating your 60 ivy bridges :p 2013-06-01 23:54:01 already done :) 2013-06-01 23:54:17 :O 2013-06-01 23:54:22 we have our own internal "fork" of alpine 2013-06-01 23:54:26 and builder 2013-06-01 23:54:39 and so, systems automatically get updates pushed out 2013-06-01 23:54:46 when the repo is updated 2013-06-01 23:54:47 :) 2013-06-01 23:55:07 I need to learn, how to build packages for alpine 2013-06-01 23:56:23 19:56:12 barthalion | Hm, any ideas why pv-grub doesn't see menu.lst of host? Path looks correct 2013-06-01 23:56:39 barthalion: what does your /etc/xen/... look like 2013-06-02 00:02:21 thanks, I will go to bed now 2013-06-02 00:02:22 bye 2013-06-02 00:02:40 bye 2013-06-02 07:24:16 let me paste relevant lines 2013-06-02 07:24:18 http://sprunge.us/jHbX 2013-06-02 07:24:20 kaniini: ^ 2013-06-02 10:03:37 barthalion: and your /boot/grub/menu.lst ? 2013-06-02 10:03:55 barthalion: also: root= should be blank on pv-grub 2013-06-02 10:04:21 oh, right, maybe that's it 2013-06-02 10:05:36 I'll test it in a moment 2013-06-02 10:06:47 root will cause extra= to be rewritten to root=/dev/xvda ro (hd0)/boot/grub/menu.lst 2013-06-02 10:07:09 which is 100% your problem probably 2013-06-02 10:15:48 kaniini: what a stupid mistake 2013-06-02 10:15:48 thanks 2013-06-02 16:53:01 algitbot: build master 2013-06-02 16:53:05 ok! 2013-06-02 16:53:10 BUILD THIS SHIT THEN 2013-06-02 16:53:40 lol wtf 2013-06-02 16:55:30 lol 2013-06-03 05:34:40 moinmoin 2013-06-03 07:19:23 good morning 2013-06-03 07:27:25 morning 2013-06-03 08:05:29 ncopa: is xen iso still concidered experimental? 2013-06-03 08:11:34 no 2013-06-03 08:11:51 clandmeter: i think we can remove "experimental" now 2013-06-03 08:17:37 ok 2013-06-03 08:22:05 ncopa: whats up with these dirs? http://sprunge.us/DLdS 2013-06-03 08:26:58 looks like i messed something up at some point 2013-06-03 08:29:00 there are also lots of checksum files which do not have an iso. I think we can trow them away. 2013-06-03 08:50:24 yeah 2013-06-03 09:16:37 files from v2.5.4-130-gd759c10 uploaded 2013-06-03 09:16:38 files from v2.4.11-30-g9f78fab uploaded 2013-06-03 09:16:39 files from v2.4.11-30-g9f78fab uploaded 2013-06-03 09:50:21 was reading about EDID, is it important for testing/linux-virt-grsec ? 2013-06-03 09:51:10 only EDID I know is about display 2013-06-03 09:51:24 so I guess that no, it isn't important :p 2013-06-03 09:54:39 util-linux may need updating, lots of changes in newer version 2013-06-03 09:56:10 yeah, but it's not so easy to upgrade, at least for me :p 2013-06-03 09:56:22 now I'm busy with Arch stuff, maybe later 2013-06-03 09:58:25 any plans for bcache, dm-cache ? 2013-06-03 10:33:35 barthalion, yes display, referring to Documentation/EDID 2013-06-03 11:56:41 Hi, please git pull https://github.com/prymeroot/aports.git Updated pekwm, live-media and libburn. 2013-06-03 12:00:55 prymeroot: please include pkgname in your commit msg's: repo/pkgname: upgrade to ... 2013-06-03 12:17:04 clandmeter: As you wish. 2013-06-03 12:45:07 found it, dm-mod.ko 2013-06-03 13:11:30 files from v2.4.11-31-g65c554a uploaded 2013-06-03 13:11:40 files from v2.4.11-31-g65c554a uploaded 2013-06-03 13:26:43 anyone know why logrotate is a depend for dovecot packages? 2013-06-03 13:26:57 likely because logging is on by default 2013-06-03 13:27:04 mm 2013-06-03 13:27:11 so it's best practice to have logrotate 2013-06-03 13:27:14 are you able to force purge packages? 2013-06-03 13:27:28 i think it is due to the shipped .logrotate file 2013-06-03 13:27:39 no if you for purge it, it will remove dovecot too 2013-06-03 13:27:39 oh 2013-06-03 13:27:43 :( 2013-06-03 13:27:55 you could just disable logrotate from running 2013-06-03 13:27:59 files from v2.3.6-240-g56b5680 uploaded 2013-06-03 13:28:01 if you don't feel the need 2013-06-03 13:28:07 i think when i installed lighttpd it created /etc/logrotate.d with its thing in there but didn't install logrotate 2013-06-03 13:28:25 it probably should have installed logrotate :) 2013-06-03 13:28:30 heh 2013-06-03 13:28:41 so you dont want logrotate? 2013-06-03 13:28:46 i clean my logs out regularly 2013-06-03 13:28:48 what do you use to rotate logs? 2013-06-03 13:28:50 no need for it personally 2013-06-03 13:28:56 you do it manually 2013-06-03 13:28:57 hm 2013-06-03 13:28:59 yep 2013-06-03 13:29:00 ;p 2013-06-03 13:29:15 i'd only save logs if there was something interesting that happened 2013-06-03 13:29:29 understand 2013-06-03 13:29:36 hm 2013-06-03 13:29:37 ok 2013-06-03 13:29:39 i have an idea 2013-06-03 13:29:45 what if we split the pkg 2013-06-03 13:29:55 hmmmm 2013-06-03 13:29:57 how would that work 2013-06-03 13:29:58 so we ship dovecot-logroate 2013-06-03 13:30:23 works for me 2013-06-03 13:30:23 :p 2013-06-03 13:30:44 which has: install_if="dovecot-$pkgver-r$pkgrel logrotate" 2013-06-03 13:30:45 that is 2013-06-03 13:30:51 apk add dovecot logrotate 2013-06-03 13:30:58 will pull in dovecot-logrotate 2013-06-03 13:31:01 is dm-cache.ko needed for dm-cache to work? 2013-06-03 13:31:14 oh cool 2013-06-03 13:31:18 vkrishn: i dont know but it sounds like it 2013-06-03 13:31:31 the drawback would be that we have many more similar 2013-06-03 13:31:35 clandmeter: commit messages fixed 2013-06-03 13:31:40 like you mentioned lighttpd 2013-06-03 13:31:49 yeah 2013-06-03 13:31:51 prymeroot: i can pull it for you if you want 2013-06-03 13:31:59 current setting , CONFIG_DM_CACHE=m 2013-06-03 13:32:10 will this produce the mod? 2013-06-03 13:32:16 vkrishn: yes 2013-06-03 13:32:37 ncopa: It would be great. 2013-06-03 13:32:56 just running a test build 2013-06-03 13:33:31 but I cannot locate the mod in v2.6-rc3 iso 2013-06-03 13:34:48 prymeroot: thanks! 2013-06-03 13:35:17 fwiw, live-media is a static lib only, so we need rebuild the pkgs that uses it to get the update 2013-06-03 13:35:38 mplayer and vlc are the only i think 2013-06-03 13:36:52 vkrishn: i found it: /lib/modules/3.9.4-0-grsec/kernel/drivers/md/dm-cache.ko 2013-06-03 13:37:00 thats current edge 2013-06-03 13:37:22 thanks 2013-06-03 13:38:28 I found some tests/benchmarks for bcache and dm-cache, having both would be nice as sometimes one is preferred over other 2013-06-03 13:48:35 http://device-mapper.com/blog/2013/04/13/benchmarking-dm-cache-with-fio/ 2013-06-03 15:25:43 I'm not on Alpine, but am trying to use Alpine's patches for building a nfs-utils against uClibc. tirpc builds fine, but rpcbind fails to build with this error: http://pastebin.com/sASeYcPC . I tried using Alpines patches for tirpc and rpcbind, but it changed nothing. Does anyone know what the solution is? 2013-06-03 20:44:43 incidentally... 2013-06-03 20:45:08 after experiencing 2013-06-03 20:45:11 nptl internals 2013-06-03 20:45:21 i want to start working on switching to musl libc 2013-06-03 20:45:36 i don't like what i see there 2013-06-03 20:45:40 at all 2013-06-03 20:48:05 so... 2013-06-03 20:48:13 i will write up what i think to alpine-devel 2013-06-03 20:48:17 the tl;dr version is this 2013-06-03 20:48:32 i am going to put the infrastructure in for an ALPINE_LIBC=musl profile 2013-06-03 20:48:45 and set up some builders for it 2013-06-03 20:48:49 and... 2013-06-03 20:48:59 we'll call it 3.0 after 2.7 is released 2013-06-03 21:13:33 n8 2013-06-04 04:41:10 kaniini: sounds good 2013-06-04 04:55:52 ncopa: any plans on mysql or we stick to oracle's 5.5? 2013-06-04 05:57:45 kaniini: nice! 2013-06-04 05:58:13 barthalion: dunno. I think we will replace mysql, but dunno when 2013-06-04 05:58:21 or to what 2013-06-04 06:08:55 Percona would be good, they care more about compatibility and ease of migration 2013-06-04 06:10:38 this is what I'd pick if we continue philosophy of providing what works more or less reliably 2013-06-04 06:26:06 bah 2013-06-04 06:45:55 moinmoin 2013-06-04 06:46:12 o/ 2013-06-04 07:02:18 morning 2013-06-04 07:17:58 rnalrd: ok if i push libvirt-1.0.6? 2013-06-04 07:19:16 yes pls 2013-06-04 08:22:48 build64-2-3: retry 2013-06-04 08:26:05 files from v2.3.6-240-g56b5680 uploaded 2013-06-04 09:36:32 kaniini: do you think we should push xen-4.2.2 to v2.6? 2013-06-04 09:37:12 i am considering push sync xen in v2.6 with the edge one 2013-06-04 09:54:24 ncopa: 01:52 <@kaniini> if you'd like 4.2.2 to be backported to 2.6, request it on alpine-devel and if ncopa says it is okay, i will do it :P 2013-06-04 09:54:52 ah ok 2013-06-04 09:54:57 i think it would be ok 2013-06-04 12:57:44 rnalrd: maybe we should pick that iscsi fix for v2.6.1? 2013-06-04 12:57:53 k 2013-06-04 12:58:07 i'm backporting it 2013-06-04 13:29:01 Just wondering, when is Alpine set to switch to musl? 2013-06-04 13:32:54 probably after 2.7, but nothing is certain yet afaik 2013-06-04 13:33:21 Heartmender: ^ 2013-06-04 13:33:54 ah, it was 2.7, not 2.6 2013-06-04 13:33:58 thanks 2013-06-04 13:34:18 oh, do you know if I can get quassel core running in alpine? 2013-06-04 13:34:34 I guess so 2013-06-04 13:34:53 quassel-core requires qt, not kdelibs if nothing has changed 2013-06-04 13:35:13 I can try to package it soon, I need to clean up some mess in Arch first 2013-06-04 13:39:04 if I can use it, I don't need debian on my server anymore 2013-06-04 13:42:11 iv'e setup a quassel core package some days ago 2013-06-04 13:42:38 so i should send it to alpine-dev to be integrated in the repo... ;-) 2013-06-04 13:43:24 barthalion: interested? 2013-06-04 13:46:56 StarWarsFan: send us a git patch (or git repo we can fetch from) and we will add it. 2013-06-04 13:47:42 exactly 2013-06-04 13:50:02 ok :) 2013-06-04 13:53:07 now I just need to find where my core data is >_> 2013-06-04 13:56:41 StarWarsFan: was quassel complicated to build? 2013-06-04 13:57:01 ncopa: not really 2013-06-04 13:57:22 just second, i'll take a look... 2013-06-04 13:57:25 can you sprunge or dpaste the apk build? 2013-06-04 13:57:37 or sprunge the git commit even better 2013-06-04 13:57:46 so author etc gets correct 2013-06-04 13:57:58 sure 2013-06-04 13:59:27 didn't use sprunge before 2013-06-04 13:59:56 so how to use it easiest way? 2013-06-04 14:00:05 apk add sprunge ; sprunge < 0001-mycommit.patch 2013-06-04 14:00:08 or 2013-06-04 14:00:11 git diff | sprunge 2013-06-04 14:00:21 nice, will try 2013-06-04 14:00:21 thx 2013-06-04 14:00:23 its a trivial wrapper script 2013-06-04 14:00:49 it will output an url 2013-06-04 14:00:52 which you can paste here 2013-06-04 14:18:04 well, the easiest way is to use curl 2013-06-04 14:20:15 sry, will take some time. some other things to do here @work at the moment... 2013-06-04 14:24:29 sure 2013-06-04 16:48:04 ncopa_: ah, btw, I hope it's ok to backport minor releases like 2.7.4 → 2.7.5 to older Alpines 2013-06-04 17:00:57 puf, arch stuff done 2013-06-04 17:01:34 Heartmender: do you need anything beyond quassel-core? 2013-06-04 17:15:57 barthalion: Nope 2013-06-04 18:34:35 sd 2013-06-04 18:34:47 wrong window :) 2013-06-04 19:09:37 does exist a motives why cyrus-sasl is at 2.1.23 version when 2.1.25 exists from 2011 ? 2013-06-04 19:10:50 probably no 2013-06-04 19:10:59 ok 2013-06-04 19:11:02 take a look at it if you have time, it may require rebuild 2013-06-04 19:11:09 i also need gssapi 2013-06-04 19:11:16 just looking at it now 2013-06-04 19:11:29 actually 2.1.26 is available 2013-06-04 19:11:51 right 2013-06-04 19:12:10 wanna look @ changelog 2013-06-04 19:12:32 nothing interesting, I did similiar upgrade in Arch month ago 2013-06-04 19:12:36 aaah 2013-06-04 19:12:44 Added support for GSS-SPNEGO SASL mechanism (Unix only), 2013-06-04 19:12:51 that's what i was looking to 2013-06-04 19:13:00 I can give you our rebuild list, but you need to grep aports anyway 2013-06-04 19:13:09 umh 2013-06-04 19:13:31 ok 2013-06-04 19:13:35 not very much packages 2013-06-04 19:13:50 but there0s libvirt 2013-06-04 19:13:50 grep for libsasl too 2013-06-04 19:14:13 nothing in main 2013-06-04 19:14:25 and in testing as well 2013-06-04 19:16:42 http://sprunge.us/jibF 2013-06-04 19:16:58 but that's not everything 2013-06-04 19:17:07 is waht is in main 2013-06-04 19:17:09 you need to rebuild libpurple and openldap too 2013-06-04 19:17:09 *what 2013-06-04 19:17:13 in main and testing 2013-06-04 19:17:15 yes, saw it 2013-06-04 19:17:31 memcached probably either 2013-06-04 19:17:39 postfix, samba, squid, svn 2013-06-04 19:17:55 that's why cyrus-sasl is still at 2.1.23 :) 2013-06-04 19:18:10 it's old because someone was lazy :p 2013-06-04 19:18:48 2.1.26 only got new symbols and some warnings, it's not rocket science 2013-06-04 19:19:21 and I see that our apkbuild is clean… 2013-06-04 19:19:28 Arch one was real mess 2013-06-04 19:19:47 a lot of patches to review and remove or keep 2013-06-04 19:22:44 with new cyrus-sasl do i need to bump pkgrel? 2013-06-04 19:23:08 i'm referring to this http://sprunge.us/jibF 2013-06-04 19:23:25 yup 2013-06-04 19:23:37 k 2013-06-04 20:38:58 bah 2013-06-04 20:39:01 libvirtd 2013-06-04 20:39:02 fails 2013-06-04 20:39:04 :/ 2013-06-04 20:44:10 weird, it shouldn't 2013-06-04 20:44:32 maybe compiler treats warnings as errors, I don't see other explanation 2013-06-04 20:44:49 In file included from rpc/virnetsaslcontext.h:24:0, 2013-06-04 20:44:49 from rpc/virnetsaslcontext.c:25: 2013-06-04 20:44:50 /usr/include/sasl/sasl.h:230:38: error: unknown type name 'size_t' 2013-06-04 20:44:50 /usr/include/sasl/sasl.h:235:5: error: unknown type name 'sasl_realloc_t' 2013-06-04 20:44:50 CC libvirt_driver_phyp_la-phyp_driver.lo 2013-06-04 20:44:50 cc1: warning: unrecognized command line option "-Wno-unused-command-line-argument" [enabled by default] 2013-06-04 20:44:53 make[3]: *** [libvirt_net_rpc_la-virnetsaslcontext.lo] Error 1 2013-06-04 20:44:57 make[3]: *** Waiting for unfinished jobs.... 2013-06-04 20:44:59 make[3]: Leaving directory `/home/fcolista/aports/main/libvirt/src/libvirt-1.0.6/src' 2013-06-04 20:45:01 make[2]: *** [all] Error 2 2013-06-04 20:45:03 make[2]: Leaving directory `/home/fcolista/aports/main/libvirt/src/libvirt-1.0.6/src' 2013-06-04 20:45:05 make[1]: *** [all-recursive] Error 1 2013-06-04 20:45:07 make[1]: Leaving directory `/home/fcolista/aports/main/libvirt/src/libvirt-1.0.6' 2013-06-04 20:45:09 make: *** [all] Error 2 2013-06-04 20:45:11 >>> ERROR: libvirt: all failed 2013-06-04 20:45:13 >>> libvirt: Uninstalling dependencies... 2013-06-04 20:45:49 I'll take a look tommorow, now I'm using my shitty arm netbook 2013-06-04 21:41:35 ACTION need some sleep... 2013-06-04 21:41:40 n8@all 2013-06-05 04:16:36 gm 2013-06-05 05:54:21 morning 2013-06-05 05:54:31 ncopa_: ah, btw, I hope it's ok to backport minor releases like 2.7.4 → 2.7.5 to older Alpines 2013-06-05 05:54:41 it is ok as long as nothing breaks ;) 2013-06-05 05:56:15 fcolista: try this maybe: http://pkgs.fedoraproject.org/cgit/cyrus-sasl.git/tree/cyrus-sasl-2.1.26-size_t.patch 2013-06-05 05:56:35 thx ncopa 2013-06-05 05:56:43 good morning, btw :) 2013-06-05 05:57:00 gm :) 2013-06-05 05:57:05 does it break abi compat? 2013-06-05 05:57:17 it's something broken with my repo, or acpica has wrong download link? 2013-06-05 05:57:25 abi compat? 2013-06-05 05:57:29 Umh...how can i check? 2013-06-05 05:57:36 checkapk 2013-06-05 05:58:01 if it only adds new symbols (but does not remove any) it should be ok to not rebuild the packages 2013-06-05 05:58:27 other things that may break abi compat if structs changes in the .h files 2013-06-05 05:59:18 checkapk directly in the dir of the apkbuild uses the latest apk built? 2013-06-05 06:00:51 http://pastebin.com/PiYUwUBC 2013-06-05 06:01:25 Downloading cyrus-sasl-2.1.23-r15 2013-06-05 06:01:25 ERROR: Unable to download 'cache/x86/cyrus-sasl-2.1.23-r15.apk' 2013-06-05 06:01:30 yes 2013-06-05 06:01:33 from cache 2013-06-05 06:01:35 yes it uses the apk in the current dir 2013-06-05 06:01:37 apk update 2013-06-05 06:02:12 or rebuild the index for cache/ 2013-06-05 06:02:16 it change soname 2013-06-05 06:02:21 ok 2013-06-05 06:02:29 buildenv-edge:~/aports/main/cyrus-sasl$ cat /etc/apk/repositories 2013-06-05 06:02:29 /home/fcolista/.cache/abuild/main 2013-06-05 06:02:29 http://dl-3.alpinelinux.org/alpine/edge/main 2013-06-05 06:02:29 then it breaks abi compat 2013-06-05 06:02:29 /home/fcolista/.cache/abuild/testing 2013-06-05 06:02:29 http://dl-3.alpinelinux.org/alpine/edge/testing 2013-06-05 06:02:30 http://upstream-tracker.org/versions/cyrus-sasl.html 2013-06-05 06:02:48 wow nice site 2013-06-05 06:03:26 wow :) 2013-06-05 06:04:47 very useful 2013-06-05 06:05:02 heh 2013-06-05 06:05:18 2012 had a peak in unintentional abi breakages 2013-06-05 06:05:28 not every lib is available, but it's easy to check if some popular library breaks other packages 2013-06-05 06:06:20 http://upstream-tracker.org/versions/icu4c.html that's the best 2013-06-05 06:06:31 ncopa, i'm trying to rebuild libvirt with cyrus-sasl patched with your link 2013-06-05 06:06:50 they keep same soname, but releases aren't backward compatible 2013-06-05 06:07:26 and poppler bumps sonames without reason 2013-06-05 06:07:37 i hate poppler for that 2013-06-05 06:08:18 icu4c breaks abi without bumping soname? 2013-06-05 06:08:43 51.2 is 99,9% backward compatible 2013-06-05 06:09:01 in Arch it broke only libreoffice 2013-06-05 06:09:36 LayoutEngine::reset 2013-06-05 06:09:49 no 2013-06-05 06:09:59 bah 2013-06-05 06:10:21 http://upstream-tracker.org/compat_reports/icu4c/51.1_to_51.2/abi_compat_report.html#Type_Problems_High 2013-06-05 06:10:38 i can see that that broke libreoffice 2013-06-05 06:10:52 not first time 2013-06-05 06:12:39 ncopa with cyrus-sasl patched, libvirt builds correctly 2013-06-05 06:12:44 nice 2013-06-05 06:13:19 barthalion: btw, it is not ok to update libs in stable branches that breaks abi 2013-06-05 06:14:46 now, grepping cyrus-sasl in main to check depedencies, all the packages has been bumped and build correctly 2013-06-05 06:15:04 in testing, kdepimlibs fails 2013-06-05 06:15:08 (only) 2013-06-05 06:15:23 but problem is a broken link download 2013-06-05 06:15:29 not cyrus-sasl 2013-06-05 06:15:51 broken link? to what? 2013-06-05 06:16:12 4.6.3 2013-06-05 06:16:16 kdepimlibs 2013-06-05 06:16:38 well, we can probably update it to 4.7 2013-06-05 06:16:50 there's 4.10 available 2013-06-05 06:16:53 oops 2013-06-05 06:17:01 4.6 no longer exists 2013-06-05 06:17:04 ok, ok, I don't use KDE :p 2013-06-05 06:17:14 4.10.3? 2013-06-05 06:17:15 dunno who's using kde with alpine :) 2013-06-05 06:17:18 4.10.4 2013-06-05 06:17:29 fair enough 2013-06-05 06:17:55 http://pkgs.fedoraproject.org/repo/pkgs/kdepimlibs/kdepimlibs-4.6.3.tar.bz2/0b0fc8d9fea0a6f26e705460a2865e18/ 2013-06-05 06:18:23 is not the right link to download the package :) 2013-06-05 06:18:41 isn't better proceeding to upgrade the package with a right link? 2013-06-05 06:19:39 http://pkgs.fedoraproject.org/repo/pkgs/kdepimlibs/kdepimlibs-4.6.3.tar.bz2/0b0fc8d9fea0a6f26e705460a2865e18/kdepimlibs-4.6.3.tar.bz2 but this one is correct 2013-06-05 06:19:58 upgrade of kdepimlibs will require another rebuild 2013-06-05 06:20:06 yes, 2013-06-05 06:20:16 taht's what is refraining me 2013-06-05 06:20:40 and probably more upgrades 2013-06-05 06:20:55 kdenetwork, kdeadmin, which will pull more packages to rebuild 2013-06-05 06:21:04 yes, lot of things.. 2013-06-05 06:21:10 ok 2013-06-05 06:21:20 i'm going with fedora link 2013-06-05 06:22:51 dont bother the kde stuff 2013-06-05 06:22:52 nah 2013-06-05 06:22:56 i think all of them are broke 2013-06-05 06:23:00 i thik so 2013-06-05 06:23:10 this wouldn't compile 2013-06-05 06:23:23 ERROR: phonon-dev: Trying to overwrite usr/share/qt/mkspecs/modules/qt_phonon.pri owned by phonon. 2013-06-05 06:23:23 ERROR: /home/fcolista/.cache/abuild/testing/x86/phonon-dev-4.5.0-r0.apk: Operation not permitted 2013-06-05 06:23:35 updating kde is a big project 2013-06-05 06:23:48 why this doesn't happens with build bot? 2013-06-05 06:23:59 it would happen if it tried to rebuild 2013-06-05 06:24:08 those pkgs have not been updated for a while 2013-06-05 06:24:11 ah 2013-06-05 06:24:15 umh. 2013-06-05 06:24:34 i think there are much broken stuff in testing now 2013-06-05 06:25:00 yeah, we need to clean it up someday 2013-06-05 06:25:08 move things to main or drop 2013-06-05 06:25:16 yeah 2013-06-05 06:25:29 we can start from kde :p 2013-06-05 06:25:44 k 2013-06-05 06:25:44 actually 2013-06-05 06:25:45 so 2013-06-05 06:25:52 kde is possible 2013-06-05 06:25:54 gnome3 is not 2013-06-05 06:26:07 so it might be worth cleaning up kde 2013-06-05 06:26:19 but it's broken and no one wants to maintain it as it deservers 2013-06-05 06:26:29 true 2013-06-05 06:26:29 deserves* 2013-06-05 06:27:41 so let's drop it and think about bringing it back later :p 2013-06-05 06:28:22 ok 2013-06-05 06:28:30 actually upstream froze some apps as LTS, so it may be worth considering, as it shouldn't require much work 2013-06-05 06:28:36 http://pastebin.com/hkx3Kgm7 2013-06-05 06:28:48 those are git status and git diff of packages 2013-06-05 06:29:46 looks ok 2013-06-05 06:36:00 i'm just worried about kdepimlibs 2013-06-05 06:36:05 shoudl we disable it? 2013-06-05 06:36:19 or i leave it untouched? 2013-06-05 06:38:19 i leave it untouched. 2013-06-05 06:42:38 I'll remove it sooner or later 2013-06-05 06:42:47 just fixing it 2013-06-05 06:43:17 it has been a wrong commit 2013-06-05 06:43:47 upgrading to 20130517 is compiling 2013-06-05 06:45:51 barthalion, looks that acpica changed licence 2013-06-05 06:45:56 now is GPL 2013-06-05 06:46:16 changes.txt:MODULE_LICENSE("GPL") to all modules. 2013-06-05 06:47:29 moinmoin 2013-06-05 06:49:16 acpica need some cleaning. 2013-06-05 06:53:26 algitbot: build master 2013-06-05 06:54:01 fcolista: ftp://ftp.cyrusimap.org/cyrus-sasl/cyrus-sasl-2.1.26.tar.gz 2013-06-05 06:54:15 barthalion, 2013-06-05 06:54:17 yes 2013-06-05 06:54:19 just one sec 2013-06-05 06:56:13 when package are in cache, if the link is broken it can be misleading 2013-06-05 06:56:30 would be a best practice clean the apk cache? 2013-06-05 06:56:48 I just remove ~/.cache/abuild 2013-06-05 06:57:02 with abuild cleancache ? 2013-06-05 06:57:08 or rm -rf ? 2013-06-05 07:00:27 files from v2.4.11-32-g39ad8b3 uploaded 2013-06-05 07:00:35 files from v2.5.4-131-g4ccb2ce uploaded 2013-06-05 07:00:36 files from v2.4.11-32-g39ad8b3 uploaded 2013-06-05 07:05:52 files from v2.3.6-241-gb788ec6 uploaded 2013-06-05 07:11:36 fyi: i think to be able to create the patch for quasselcore later today 2013-06-05 07:11:49 this was my first setup of an alpine package 2013-06-05 07:12:13 so i have to do some changes/cleanup to make it fit into the system 2013-06-05 07:12:16 fcolista: rm -rf, I'm a simple man 2013-06-05 07:12:26 k barthalion 2013-06-05 07:41:28 ncopa, i'm using with success apache with mod-auth-kerb 2013-06-05 07:41:37 can be moved to main? 2013-06-05 07:41:41 sure 2013-06-05 07:41:51 k 2013-06-05 07:49:48 question: do we _really_ use "unstable" repository? 2013-06-05 07:50:00 isn't testing the same? 2013-06-05 07:50:17 Or we should move kde stuff that doesn't compile to unstable? 2013-06-05 07:50:39 we dont use unstable afaik 2013-06-05 07:51:04 i think we could have an unmaintained subdir instead of unstable 2013-06-05 07:51:13 +1 2013-06-05 07:51:19 have more sense 2013-06-05 07:51:43 we ould rename 'unstable' to 'unmaintained' 2013-06-05 07:51:51 and move things from testing thats broke to there 2013-06-05 07:51:57 right 2013-06-05 07:51:58 if someone wants to pick it up, its there 2013-06-05 07:52:26 and then, when it donesn't break anything, can be moved to testing for futher tests 2013-06-05 07:52:37 *further 2013-06-05 07:52:47 yes 2013-06-05 07:52:54 once it builds again 2013-06-05 07:55:25 exactly 2013-06-05 10:16:03 files from v2.5.4-132-g0a71931 uploaded 2013-06-05 10:16:25 files from v2.5.4-132-g0a71931 uploaded 2013-06-05 10:22:38 files from v2.3.6-242-g563e2f3 uploaded 2013-06-05 10:22:57 files from v2.3.6-242-g563e2f3 uploaded 2013-06-05 11:27:22 files from v2.4.11-33-gef7cc55 uploaded 2013-06-05 11:29:27 files from v2.4.11-33-gef7cc55 uploaded 2013-06-05 11:40:50 hi 2013-06-05 11:40:56 mirror 5 down for maintenance (updates) 2013-06-05 11:41:01 will take 20 minutes or so 2013-06-05 11:47:30 ok, thanks for the notification 2013-06-05 11:56:02 back up again :) 2013-06-05 11:56:35 that was quick :) 2013-06-05 11:57:01 yeah just updated the host 2013-06-05 12:08:32 hm 2013-06-05 12:08:40 the linux kernel is near 3.11 2013-06-05 12:08:46 that means it gets network support soon ;) 2013-06-05 12:08:54 or are they going for 32bit and jump to 4.0? 2013-06-05 12:10:02 lol 2013-06-05 12:27:34 re 2013-06-05 13:23:50 build-2-3: retry 2013-06-05 13:23:54 build64-2-3: retry 2013-06-05 13:37:32 files from v2.3.6-243-g9da25b8 uploaded 2013-06-05 13:40:46 files from v2.3.6-243-g9da25b8 uploaded 2013-06-05 14:02:56 files from v2.5.4-140-gf63fa9c uploaded 2013-06-05 14:03:27 files from v2.4.11-37-g7a4dbd1 uploaded 2013-06-05 14:10:22 files from v2.5.4-141-ge466dbb uploaded 2013-06-05 14:12:20 files from v2.4.11-38-ga2883b6 uploaded 2013-06-05 14:14:19 files from v2.5.4-141-ge466dbb uploaded 2013-06-05 21:30:20 ncopa: barthalion: ok, here's the quasselcore-diff: http://sprunge.us/PQQK 2013-06-05 21:30:53 currently untested in detail 2013-06-05 21:31:14 i was able to connect to the core with a client 2013-06-05 21:31:53 but this was on a 2.4.9-release of alpine 2013-06-05 21:32:28 so on 2.6.x it must be tested... 2013-06-05 21:32:45 please let me know if i should change something... 2013-06-05 21:32:52 and now: n8@all :-) 2013-06-06 04:30:02 Moinmoin 2013-06-06 04:55:51 huomenta 2013-06-06 06:26:55 god morgen 2013-06-06 07:03:15 how should be handled the chown | chgroup in $pkgname.post-upgrade script? abuild complains that it has chown/chgroup 2013-06-06 07:03:48 why do you need it in post-upgrade? 2013-06-06 07:04:05 duno, i't ejabberd 2013-06-06 07:04:12 which has this 2013-06-06 07:04:13 you should normally set permissions in .apk itself 2013-06-06 07:04:18 in package() or similar 2013-06-06 07:04:37 ah 2013-06-06 07:05:03 you should add pkgusers and pkggroups in APKBUILD 2013-06-06 07:05:08 Done 2013-06-06 07:05:18 but this applies to install, right? 2013-06-06 07:05:27 and you chsoul chown the dirs from there 2013-06-06 07:05:28 Implicitly applies also to post-upgrade 2013-06-06 07:05:29 yes 2013-06-06 07:05:44 and it also works for upgrades 2013-06-06 07:05:46 so i can remove post-upgrade, basically 2013-06-06 07:05:48 ok 2013-06-06 07:05:51 yes 2013-06-06 07:05:58 i needed for this confirm 2013-06-06 07:05:59 thx 2013-06-06 08:04:10 StarWarsFan: I'll take a look soon 2013-06-06 08:59:21 ncopa: do you mind if i add https://git.gnome.org/browse/vte/patch/?id=b73782a28894e25ed146271f9d6c6775a6836199 to vte? 2013-06-06 08:59:47 wha does it do? why? 2013-06-06 09:00:11 why do you need it? 2013-06-06 09:00:21 i dont mind, im just curious why you want it 2013-06-06 09:00:24 to make alt+1/2/3 2013-06-06 09:00:33 work in terminal 2013-06-06 09:00:38 aha 2013-06-06 09:00:44 i think i want that too :) 2013-06-06 09:00:45 i cannot switch inside irssi now 2013-06-06 09:01:01 yes apply it please 2013-06-06 09:01:12 explain why it was needed in commit msg 2013-06-06 09:01:41 ok 2013-06-06 09:04:38 Cloudcity 2013-06-06 09:36:34 ncopa: can i upgrade a pacakge with repo pinning? 2013-06-06 09:37:31 i suppose with apk add -u package@repo 2013-06-06 09:38:58 that doesnt do it 2013-06-06 09:39:14 it adds it to world, but doesnt do that upgrade 2013-06-06 09:39:25 try apk del the pkg ? 2013-06-06 09:39:28 that/the 2013-06-06 09:40:02 sorry.. i'm messing with my new irc client, hexchat 2013-06-06 09:40:46 ncopa: i cannot del that pkg, its pulled in with alpine-desktop 2013-06-06 09:41:14 then i dont know :-/ 2013-06-06 09:47:16 ha works :) 2013-06-06 09:48:42 ncopa: could we push that to 2.6? 2013-06-06 09:48:55 vte? 2013-06-06 09:49:00 yes 2013-06-06 09:50:12 ok, i can do it 2013-06-06 09:50:22 ncopa: don't hurry, just a short question: you've seen my posting from yesterday evening? 2013-06-06 09:50:56 about quassel-core? 2013-06-06 09:51:11 I did, gimme sec :p 2013-06-06 09:51:44 StarWarsFan: i saw it and i saw that barthalion is working on it 2013-06-06 09:51:52 ah ok, nice 2013-06-06 09:51:53 thx 2013-06-06 09:51:56 :-) 2013-06-06 09:52:16 tere was other patch we talked about yesterday.. 2013-06-06 09:52:20 ah cfengine 2013-06-06 09:52:48 ncopa: I would like to maintain apk module for Ansible in git.alpinelinux.org, what should I do to create a repo? 2013-06-06 09:54:57 git clone --bare ansible ansible.git ; scp -r ansible.git git.alpinelinux.org:cgit/ 2013-06-06 09:55:03 i think that should do it 2013-06-06 09:55:21 i suppose you should edit ansible.git/description before you scp it 2013-06-06 09:55:24 create cgitrc 2013-06-06 09:55:31 not needed anymore... 2013-06-06 09:55:37 oh ok 2013-06-06 09:55:48 thanks, I'll try after I push quassel-core 2013-06-06 09:56:32 clandmeter: without any wgeetrc it will use $USER as section 2013-06-06 09:56:49 as i see it right now: 2013-06-06 09:57:30 i named it "quasselcore" as this is the used name on quassel-irc.org 2013-06-06 09:57:37 i mean, without "-" 2013-06-06 09:57:52 eah, but someday, someone can want to push client 2013-06-06 09:58:10 so it will be cleaner to use quassel-core IMHO 2013-06-06 09:58:35 maybe that's me? ;-) 2013-06-06 09:58:39 yeah, why not... 2013-06-06 10:00:46 i think we shoudl have a pkg 'quassel' then a subpkg in tehre named 'core' 2013-06-06 10:01:18 if we build a client later we make another subpkg named -client 2013-06-06 10:02:14 then I will package both :p 2013-06-06 10:03:35 ACTION is happy with irssi working like normal again 2013-06-06 10:12:38 hm 2013-06-06 10:12:52 I'm not sure if I understand documentation correctly 2013-06-06 10:12:54 subpackages="$pkgname-core:core $pkgver-client:client" 2013-06-06 10:13:02 And then I define function package_core and package_client? 2013-06-06 10:13:06 functions* 2013-06-06 10:13:12 just core is enough 2013-06-06 10:13:44 thanks 2013-06-06 10:14:11 but for $pkgname subpkgs you dont need to use :funct afaik 2013-06-06 10:15:01 but I use hyphen in subpackages' names 2013-06-06 10:15:05 ie: $pkgname-client is enough for client() {} 2013-06-06 10:15:11 och 2013-06-06 10:15:21 yes if you use special chars you need it 2013-06-06 10:16:46 is there a plan to migrate on mariadb any time soon??? 2013-06-06 10:17:22 not yet, and there was no decision to pick MariaDB 2013-06-06 10:19:54 thnx barthalion 2013-06-06 10:22:29 barthalion: i didnt understand your question. its pkgname:funct where pkgname can be anything and funct is limited 2013-06-06 10:22:57 clandmeter: but I understood your answer :p 2013-06-06 10:48:01 StarWarsFan: this is what I have so far: http://sprunge.us/CaQH 2013-06-06 10:52:17 i'll take a look 2013-06-06 10:53:46 for some reason it fails to install client, but we can just disable it 2013-06-06 10:53:56 and problem solved :p 2013-06-06 10:54:44 barthalion: looks ok to me 2013-06-06 10:54:56 but i'm more or less a newbie on alpine ;-) 2013-06-06 10:55:32 ACTION <- a lot to learn he has *g* 2013-06-06 10:55:42 :D 2013-06-06 10:55:59 especially the part with sub-packages is completely new for me 2013-06-06 10:56:09 very interesting! 2013-06-06 10:56:15 you're not alone, it looks different in Arch 2013-06-06 11:06:22 hm, now I need to cheat abuild to ignore quassel itself 2013-06-06 11:06:25 ncopa: ^ 2013-06-06 11:06:26 ideas? 2013-06-06 11:06:55 relevant log http://sprunge.us/NDPS 2013-06-06 11:08:55 another question: what's the common way for contributor/maintainer entry? 2013-06-06 11:08:59 nick or real name? 2013-06-06 11:10:02 looks like real name, right? 2013-06-06 11:10:20 it's up to you 2013-06-06 11:11:15 so i think to use it as most of the already existing packages 2013-06-06 11:11:21 s/starwarsfan/Yves Schumann/g 2013-06-06 11:11:23 ;-) 2013-06-06 11:13:17 barthalion: i'd do make install, then in subpackage split function i'd mv "$pkgdir"/* "$subpkgdir"/ 2013-06-06 11:20:39 when I manage to get quassel-client to work, pkg quassel will be metapackage depending on both -core and -client 2013-06-06 11:20:43 now it is, because it is 2013-06-06 11:21:41 Heartmender: ^ 2013-06-06 11:21:49 StarWarsFan: ^ 2013-06-06 11:21:54 :-) 2013-06-06 12:16:15 virtinst depends on libvirt, which depends on cyrus-sasl 2013-06-06 12:17:38 will virtinst need rebuild too ? 2013-06-06 12:18:15 yes...it's enough bumping pkgrel 2013-06-06 12:20:21 er, why? 2013-06-06 12:20:33 it doesn't matter for virtinst 2013-06-06 12:20:48 if it is among the depends, it needs rebuild 2013-06-06 12:21:07 yeah, but virtinst doesn't depend on cyrus-sasl 2013-06-06 12:21:08 or libsasl 2013-06-06 12:21:11 ah 2013-06-06 12:21:13 ok 2013-06-06 12:22:14 yes, i misread: vkrishn is referring to a virtinst depends on libvirt 2013-06-06 12:22:26 that in turns depends from cyrus-sasl 2013-06-06 12:22:26 ok 2013-06-06 12:22:39 iirc, kyotocabinet is improved version over tokyocabinet 2013-06-06 12:22:51 no vkrishn, it's enough rebuilding libvirt 2013-06-06 12:22:57 which has be done 2013-06-06 12:22:59 ok, 2013-06-06 12:23:10 ok 2013-06-06 12:42:32 ncopa: I scped bare repo to ~/cgit, could you make it visible? 2013-06-06 12:43:35 should show up auto 2013-06-06 12:44:15 ah 2013-06-06 12:44:18 symlink is missing 2013-06-06 12:45:15 showed up now 2013-06-06 12:45:38 cool, thanks 2013-06-06 12:45:45 eh 2013-06-06 12:45:51 the description 2013-06-06 12:46:01 you should scp it to the bare repo 2013-06-06 12:46:02 oops 2013-06-06 12:46:32 it corresponds to .git/description in the checked out repo 2013-06-06 12:47:06 scp description git.alpinelinux.org:cgit/ansible-apk.git/ 2013-06-06 12:47:17 yeah, fixed 2013-06-06 12:47:43 you can alos add a README.md if you want 2013-06-06 12:47:55 that shuold be committed 2013-06-06 12:48:21 sure 2013-06-06 12:48:29 as soon as I write it :p 2013-06-06 12:50:10 hm, owner is "Linux User" 2013-06-06 12:50:23 I guess I should set something on git.ao.o 2013-06-06 13:39:27 files from v2.5.4-142-g5219257 uploaded 2013-06-06 14:41:22 hi 2013-06-06 14:41:29 hi 2013-06-06 14:42:35 I have a problem: While I am connected with the Intel AMT KVM, I can not boot up alpine from an usb stick "SYSLINUX 5.02 EDD Load error - Boot error" 2013-06-06 14:43:00 If I am not connected with the Intel AMT KVM, it works well?! 2013-06-06 14:44:56 ? 2013-06-06 14:45:01 sounds strange 2013-06-06 14:46:25 I think, it has something to do with the message before "Waiting Up to 8 Minutes For KVM FW ....." 2013-06-06 14:46:41 i dunno... 2013-06-06 14:47:02 ok :S 2013-06-06 14:48:16 ncopa: could we add usb keyboard support to initfs? difficult to debug in recovery shell without keyboard. 2013-06-06 14:48:35 i thought it already was? 2013-06-06 14:48:58 not with the keyboard i had recently 2013-06-06 14:49:47 hm 2013-06-06 14:49:47 add it 2013-06-06 14:49:47 do you know which kernel module it is? 2013-06-06 14:50:00 hmm let me check 2013-06-06 14:55:58 usbkbd 2013-06-06 14:57:05 i thought that was in the initramfs already 2013-06-06 15:02:06 maybe its just this keyboard. i cannot reproduce it now. 2013-06-06 15:02:18 im quite sure it doesnt work at home. 2013-06-06 17:47:37 ncopa: I think, I need to use the stupid mode of syslinux 2013-06-06 17:47:50 syslinux -s 2013-06-06 20:04:28 ncopa: barthalion: still here? 2013-06-06 20:04:51 ACTION <- playing with quassel-core package 2013-06-06 20:05:05 which is not working at mo 2013-06-06 20:05:10 so i have some questions 2013-06-06 20:09:16 sure 2013-06-06 20:09:30 I should have tested it, but well 2013-06-06 20:16:40 will create a patch 2013-06-06 20:16:46 because some path's are wrong now 2013-06-06 20:16:58 after renaming it from quasselcore to quassel-core 2013-06-06 20:17:06 but there are another bigger issue 2013-06-06 20:17:28 core crashes during connect of a client :-( 2013-06-06 20:17:46 uh-uh 2013-06-06 20:17:59 activating debug log right now and check what's going on... 2013-06-06 20:18:07 great 2013-06-06 20:18:26 drop me a line if you find out why it's broken 2013-06-06 20:18:37 sure 2013-06-06 20:24:39 hm, bad :-( 2013-06-06 20:24:45 no log output 2013-06-06 20:24:50 only result: 2013-06-06 20:25:04 /etc/init.d/quassel-core status returns "crashed" 2013-06-06 20:25:11 any ideas how to dig? 2013-06-06 20:25:36 nothing on /var/log/messages and nothing on dmesg so far... 2013-06-06 20:27:24 you should probably recompile it with -O0 -g 2013-06-06 20:27:27 and use gdb 2013-06-06 20:28:04 I'll try to debug it tommorow 2013-06-06 20:28:28 ok great 2013-06-06 20:28:38 i create a patch for the other issues 2013-06-06 20:29:00 cool, send it to alpine-devel or to my e-mail 2013-06-06 20:29:08 thanks 2013-06-06 20:32:36 this is the point to configure git send-email :-) 2013-06-06 20:39:04 I never got through it, my config looked correct but send every message twice 2013-06-06 20:39:39 that's the main reason because I sat here and asked for pull request 2013-06-06 20:41:20 err, I have no idea why I wrote "because" 2013-06-06 20:41:27 I need more sleep :p 2013-06-06 20:41:38 *g* 2013-06-06 20:41:41 ACTION too... 2013-06-06 20:41:58 goodnight 2013-06-06 21:56:55 btrfs-progs is too old?! 2013-06-06 21:56:57 v0.19? 2013-06-06 22:05:50 barthalion: patch done, see ML :-) 2013-06-06 22:22:22 anyone feel like upgrading prosody to 0.9 rc2? :P 2013-06-06 22:22:28 includes ipv6 support 2013-06-07 04:33:14 Moinmoin 2013-06-07 06:38:31 I did some refactoring of quassel apkbuild 2013-06-07 06:38:52 quassel is pretty bloated imho 2013-06-07 07:01:07 ncopa: is the dependency to sqlite added during build? 2013-06-07 08:54:36 sqlite? dunno 2013-06-07 08:54:59 i've added this in my patch yesterday 2013-06-07 08:55:10 but seems to be wiped out during your refactoring 2013-06-07 08:56:33 ah 2013-06-07 08:56:43 i didnt check my inbox 2013-06-07 08:56:44 sorry 2013-06-07 08:56:47 so 2013-06-07 08:56:52 is it a runtime dependency? 2013-06-07 08:56:57 yes 2013-06-07 08:57:01 it needs the sqlite cmd line util? 2013-06-07 08:57:13 or it needs the libs? 2013-06-07 08:57:14 oh, that's a good question 2013-06-07 08:57:19 i think the libs 2013-06-07 08:57:43 during first connect using a client, quassel gets configured 2013-06-07 08:57:55 during these steps you could choose the used database 2013-06-07 08:58:20 so:libQtSql.so.4 2013-06-07 08:58:47 i think what you need to do is manually install qt-sqlite 2013-06-07 08:58:57 thats what i suspect 2013-06-07 08:59:25 what do you mean with "manually"? 2013-06-07 08:59:34 apk add qt-sqlite 2013-06-07 08:59:40 i think that should be enough 2013-06-07 08:59:49 and why not as a dependency? 2013-06-07 09:00:19 because i suspect you could chose mysql or pgsql too, incase you dont want use sqlite 2013-06-07 09:00:41 right, you can choose between pgsql and sqlite 2013-06-07 09:00:54 if we add it as hard dep, you'll always end up with having sqlite there 2013-06-07 09:01:43 hm, wouldn't be a problem for me... 2013-06-07 09:01:58 i suppose it doesnt matter 2013-06-07 09:01:59 so $USER must see a hint to install proper db 2013-06-07 09:02:20 quassel is soo bloated anyways 2013-06-07 09:02:39 i suppose it doesnt matter to add a another MB dead meat 2013-06-07 09:02:46 *g* 2013-06-07 09:02:51 and in most cases, i suppose sqlite makes more sense than pgsql 2013-06-07 09:02:57 i doubt anyone will use pgsqlite 2013-06-07 09:03:00 pgslq 2013-06-07 09:03:06 pgsql 2013-06-07 09:03:12 i cannot type 2013-06-07 09:04:23 StarWarsFan: could you please test that it is qt-sqlite that is needed and not 'sqlite'? 2013-06-07 09:04:28 (i'm just guessing) 2013-06-07 09:31:29 build-2-5: retry 2013-06-07 09:31:33 or something 2013-06-07 09:41:18 files from v2.5.4-143-g6f4d7f2 uploaded 2013-06-07 09:47:49 ncopa: sure, i check this 2013-06-07 09:51:19 ncopa, i've a question re -git package and no-git package 2013-06-07 09:51:33 i'm going to do a new apkbuild for ejabberd, the git version 2013-06-07 09:51:45 since the new branch supports gssapi 2013-06-07 09:51:54 but not the stable 2013-06-07 09:52:18 so i'm going to create ejabberd-git, like tvheadend-git, in testing 2013-06-07 09:52:38 afaik that was tricky 2013-06-07 09:52:42 I mean undocumented :p 2013-06-07 09:52:49 when i do the git snapshot, it's ok if i name it ejabber-git ? 2013-06-07 09:52:59 is bad documented barthalion 2013-06-07 09:53:15 (if you are referring to ejabberd) 2013-06-07 09:53:17 wasn't it possible to fetch exact commit? 2013-06-07 09:53:35 no, I'm refering to vcs support in abuild 2013-06-07 09:54:02 i find it very useful 2013-06-07 09:54:09 just two options 2013-06-07 09:54:14 giturl and disturl 2013-06-07 09:54:32 and then abuild snapshot upload it on archive/$pkgname on d.a.o 2013-06-07 09:54:37 this is not my problem 2013-06-07 09:54:43 my problem is how to name it 2013-06-07 09:55:08 you misunderstood me 2013-06-07 09:55:12 the archive on d.a.o shoudl be called $pkgname-git or $pkgname ? 2013-06-07 09:55:19 abuild can download exact commit from git repository 2013-06-07 09:55:25 there is no need to make a tarball 2013-06-07 09:55:38 is not a single commit 2013-06-07 09:55:56 well, it is 2013-06-07 09:56:32 let me check in abuild code 2013-06-07 09:56:37 giturl= 2013-06-07 09:56:37 ok 2013-06-07 09:56:40 reporev= 2013-06-07 09:56:58 i mean: http://www.ejabberd.im/cyrsasl_gssapi 2013-06-07 09:57:08 you can change branch too 2013-06-07 09:57:39 "You can checkout a specific branch in git by adding -b $branch in $giturl." 2013-06-07 09:57:58 nice 2013-06-07 09:57:59 pl 2013-06-07 09:58:02 ok 2013-06-07 09:58:08 s/pl/ok 2013-06-07 09:58:21 anyway 2013-06-07 09:58:38 gssapi support is quite messed up in ejabberd 2013-06-07 09:58:55 the last patch applies to 2.1.10, but the stable version is 2.1.12 2013-06-07 09:59:20 the page states: 2013-06-07 09:59:20 Note: this patch was included in ejabberd master branch, and will be included in the future ejabberd 3.0.0-alpha releases. You can test it there. 2013-06-07 09:59:46 and it doesn't work? 2013-06-07 09:59:46 so i'm wondering that a "clean" thing is leave the stable version withouth patches 2013-06-07 10:00:01 and makes the -git package from master branch 2013-06-07 10:00:10 if the patch applies to 2.1.12, I don't see problem to bring it in 2013-06-07 10:00:25 patch applies to 2.1.10 2013-06-07 10:00:52 but it doesn't applies to 2.1.12 2013-06-07 10:02:02 I would ask patch author if it's possible to fix patch 2013-06-07 10:02:18 but if you want to use unstable version, why not :p 2013-06-07 10:02:39 :) 2013-06-07 10:02:57 since is in testing, i tought it make sense 2013-06-07 10:03:09 but if u have other advices, they are welcome 2013-06-07 10:03:13 sure, if you name it $pkgname-git 2013-06-07 10:03:22 yes. 2013-06-07 10:03:35 no problem 2013-06-07 10:03:43 So the source entry will be $pkgname-git or $pkgname ? 2013-06-07 10:03:49 That is my original question .) 2013-06-07 10:03:50 :) 2013-06-07 10:03:59 $pkgname-git 2013-06-07 10:04:00 source="http://dev.alpinelinux.org/archive/$pkgname/$pkgname-$pkgver.tar.gz 2013-06-07 10:04:16 $pkgname = ejabberd-git 2013-06-07 10:04:20 but again, you don't have to make a tarball 2013-06-07 10:04:31 as abuild can fetch sources from git 2013-06-07 10:04:48 Mp5shooter: can we wait for stable 0.9.0? I would move it to main then 2013-06-07 10:07:07 will there be a package for microcode fix intel-ucode or amd-ucode??! 2013-06-07 10:07:58 sure, if someone package it and put in non-free 2013-06-07 10:09:08 it would help alot on some machines 2013-06-07 10:09:27 I'll add it to my todo 2013-06-07 10:11:35 thnx it is small package but I really don`t know how to deal with grsec-kernel and still tryin to learn apk packaging when my job permits 2013-06-07 11:21:40 barthalion: we prefer a tarball so our buildboxes dont need to use svn/git 2013-06-07 11:22:24 the maintainer is responsible for a proper archive. 2013-06-07 11:23:11 fcolista: other option, create a patch that you call git.patch or similar 2013-06-07 11:23:12 so is there any rationale behind vcs support? tarball should be uploaded anyway 2013-06-07 11:23:33 so a dev can create one with one cmd 2013-06-07 11:23:35 fcolista: you could do foo-1.0_git 2013-06-07 11:24:08 and if you have git access for alpine, you can host it on our server 2013-06-07 11:25:10 and the abuild function will also take care of the versioning 2013-06-07 11:25:48 it will automatically do it like ncopa suggested 2013-06-07 11:26:24 in general, i think we should avoid git versions 2013-06-07 11:26:31 fcolista: only add -git if you want to have 2 packages in our repo (which most of the time doesnt make any sence) 2013-06-07 11:26:38 we need to maintain it if it slips into a release 2013-06-07 11:37:27 back from lunch 2013-06-07 11:37:59 pkgname=ejabberd-git 2013-06-07 11:37:59 _pkgname=ejabberd 2013-06-07 11:37:59 pkgver=0_git20130607 2013-06-07 11:38:17 source="http://dev.alpinelinux.org/archive/$pkgname/$pkgname-$pkgver.tar.gz 2013-06-07 11:38:31 sounds good? 2013-06-07 13:23:16 files from v2.5.4-144-gae7699b uploaded 2013-06-07 13:23:16 files from v2.5.4-144-gae7699b uploaded 2013-06-07 14:18:10 search bar broke in firefox-21 2013-06-07 14:51:26 not sure, is this http://git.kernel.org/pub/scm/linux/kernel/git/rt/ (real time) version of kernel? 2013-06-07 14:51:46 any possibility of having it in alpine? 2013-06-07 14:52:12 do you need it? 2013-06-07 14:52:25 I can package it if you provide a config 2013-06-07 14:52:28 though I guess applying grsec/pax may be not possible 2013-06-07 14:52:45 yeah, only rt patch 2013-06-07 14:52:53 barthalion, if one is using al on microdevices, it may be nice 2013-06-07 14:53:13 specially critical appliances 2013-06-07 14:53:21 like gauges 2013-06-07 14:53:32 well, I'll add it to todo 2013-06-07 14:53:45 3.2 will be ok? 2013-06-07 14:54:10 yes, thats an issue as the repo is not up to date 2013-06-07 14:54:26 https://www.kernel.org/pub/linux/kernel/projects/rt/ 2013-06-07 14:54:38 3.8 is the newest, but I don't think that I can maintain it well 2013-06-07 14:54:44 yes the rt repo, I think last branch is 3.4y 2013-06-07 14:54:59 And taking lts release as a base will save some time 2013-06-07 14:55:07 3.8 on rt, cool 2013-06-07 14:55:07 as we will need only to bump pkgver and patch version 2013-06-07 14:55:39 3.8 is unsupported 2013-06-07 14:55:48 so it's matter of choice between 3.2 and 3.4 2013-06-07 14:55:53 last I checked 3.4y 2013-06-07 14:56:09 so I'll package 3.4 2013-06-07 14:56:31 I think 3.4y , maybe in sync with vserver features 2013-06-07 14:56:56 ok nice 2013-06-07 14:56:57 I'll work on it in sunday 2013-06-07 14:57:18 thnx 2013-06-07 15:03:23 I try to find more on what config options already in al 3.4 kernel may not build with rt 2013-06-07 15:03:39 I would* 2013-06-07 15:07:05 gtg 2013-06-07 15:10:51 sure barthalion 2013-06-07 15:23:21 algitbot: build master 2013-06-07 15:23:24 OK 2013-06-07 15:27:21 algitbot: so build it 2013-06-07 15:31:20 good boy 2013-06-07 15:33:48 what use would someone have for pacman in alpine? 2013-06-07 15:34:13 same as deboostrap and rpm 2013-06-07 15:34:26 making chroots for containers and/or virtual machines 2013-06-07 15:34:31 ah 2013-06-07 16:29:49 anything on current state of uclibc on rt, 2013-06-07 16:30:37 this https://rt.wiki.kernel.org/index.php/HOWTO:_Build_an_RT-application - last revised on 2009-11-15 , says "uClibc is not recommended" 2013-06-07 16:33:09 found few more limitations, https://rt.wiki.kernel.org/index.php/Frequently_Asked_Questions 2013-06-08 04:53:47 hmm 2013-06-08 04:53:49 what is xtables 2013-06-08 04:54:03 /etc/xtables dir suddenly appeared after an upgrade 2013-06-08 22:06:19 musl is built with $ALPINE_LIBC on any setting, as musl-gcc package can be used as a bootstrap 2013-06-08 23:25:08 kaniini: Starting the transition of Alpine to Musl? 2013-06-08 23:51:48 some initial groundwork 2013-06-08 23:51:56 it allows us to do a tinderbox run 2013-06-09 00:40:27 what are you looking to burn down 2013-06-09 00:46:04 tinderbox is a test environment 2013-06-09 00:46:12 :p 2013-06-09 13:39:28 https://github.com/prymeroot/aports – updated gdk-pixbuf and cmake. Pull if you want. 2013-06-09 13:43:11 let's see if I fucked up something 2013-06-09 13:43:14 apparently no 2013-06-09 13:43:55 algitbot: build master 2013-06-09 13:46:08 prymeroot: thanks 2013-06-09 13:47:48 barthalion: you're welcome 2013-06-09 13:50:09 ncopa: can we talk about missing directories in /run and /var/run, enabling tmpfiles.d by default and providing its files? 2013-06-09 13:50:38 creating directory in init.d scripts is dirty IMO 2013-06-10 06:11:33 morning 2013-06-10 06:11:47 creating directory in init.d scripts is dirty IMO 2013-06-10 06:11:55 i think its the only way 2013-06-10 06:12:17 because /run and /var/run is most likely a tmpfs 2013-06-10 06:12:29 so after a boot no dirs are there 2013-06-10 06:34:47 anyone have opinion about enabling spice protocol for xen? 2013-06-10 06:34:55 i use it with qemu/kvm 2013-06-10 06:37:18 ncopa: Is there anything like apt-get build-dep in alpine? 2013-06-10 06:37:49 what is build-dep? 2013-06-10 06:38:05 all dependencies needed for building packages? 2013-06-10 06:38:07 Yes 2013-06-10 06:38:09 like 2013-06-10 06:38:17 apt-get build-dep znc 2013-06-10 06:38:19 etc 2013-06-10 06:38:43 what does build-dep include? 2013-06-10 06:38:57 whats the difference from build-essential? 2013-06-10 06:39:13 like libssl-dev for znc ssl 2013-06-10 06:39:15 well, it fetches all build dependencies of given package 2013-06-10 06:39:20 headers and so on 2013-06-10 06:39:34 ok 2013-06-10 06:39:37 abuild installdeps 2013-06-10 06:39:39 ncopa: it's not the only way, since some time there is tmpfiles.d support in openrc 2013-06-10 06:40:02 ncopa: abuild installdeps $package? 2013-06-10 06:40:19 Niichan: no, cd aports/main/$package && abuild installdeps 2013-06-10 06:40:51 the make dependencies are not stored in the run-time apk index (saves space) 2013-06-10 06:40:57 so you need the aports tree for that 2013-06-10 06:41:00 ncopa: Ah 2013-06-10 06:41:20 barthalion: I'll have to read on tmpfiles.d 2013-06-10 06:41:38 Niichan: there is also a meta pkg, build-base 2013-06-10 06:41:54 you will need that 2013-06-10 06:42:03 i think it corresponds to build-essentials 2013-06-10 06:42:37 and there is a meta package that pulls in abuild and git and those: alpine-sdk 2013-06-10 06:43:15 so: apk add alpine-sdk; git clone git://git.alpinelinux.org/aports; cd aports/main/$package && abuild installdeps 2013-06-10 06:44:26 abuild installdeps will install a 'virtual' meta package 2013-06-10 06:44:36 an empty package named .-makedepends 2013-06-10 06:44:55 witch depends on all the dependencies for the give pkg 2013-06-10 06:45:12 so when you uninstall it, you can simply: apk del .-makedepends 2013-06-10 11:45:54 kaniini: around? 2013-06-10 13:58:39 hmpf 2013-06-10 13:58:45 whats this: /usr/bin/python2.7: symbol '__gmpz_mul': can't resolve symbol 2013-06-10 14:02:38 i wonder if it is related gmp upgrade to 5.1.2 2013-06-10 15:54:31 clandmeter: yes 2013-06-10 15:54:46 clandmeter: i think we'll ship alpine/arm as musl 2013-06-10 16:16:08 kaniini, agreed. have you made any progress on arm? 2013-06-10 16:16:54 i have acquired a versatile express (!) 2013-06-10 16:16:56 i have a prospect on arm based thing, so i might get some interest on it within few months... 2013-06-10 16:18:08 will ping you later on this... 2013-06-10 16:18:12 gotta go for now 2013-06-10 16:20:54 my plan is to ship arm on musl only. 2013-06-10 17:51:40 kaniini: are you building alpine on musl for all archs, or just shooting for arm? 2013-06-10 17:56:31 jbilyk: the eventual plan is to get rid of uclibc across all archs, but we will not do this for 2.7. 2013-06-10 17:56:41 new archs should use musl. 2013-06-10 17:56:47 ok 2013-06-10 17:56:59 kaniini: are you shooting for arm support for 2.7? 2013-06-10 17:57:02 it is possible to run a tinderbox build with musl on any arch. 2013-06-10 17:57:09 apk add musl musl-gcc 2013-06-10 17:57:15 then set CC=musl-gcc 2013-06-10 17:57:17 :) 2013-06-10 17:57:39 cool 2013-06-10 17:57:41 nice work 2013-06-10 17:57:58 this is incomplete, but it means we can bootstrap with aports. 2013-06-10 17:58:21 without dealing with crosstool (!) 2013-06-10 18:30:02 ncopa: ah, I know why the owner of my repo is "Linux User" 2013-06-10 18:30:22 ncopa: you need to change my name on git.al.o 2013-06-10 18:36:20 hm ! 2013-06-10 18:36:30 my apk is broken 2013-06-10 18:36:36 apk add --> 100% cpu since upgrading 2013-06-10 18:41:30 kaniini: should i still get that hardkernel board? 2013-06-10 18:42:17 yes, seems reasonable if you want to go for it. 2013-06-10 18:42:27 vexpress is slow :P 2013-06-10 18:43:33 ok ill order it tmorrow 2013-06-10 18:43:57 i have a stage1 environment for arm 2013-06-10 18:44:14 and musl :) 2013-06-10 18:45:13 did you use sabotage? 2013-06-10 18:53:02 no 2013-06-10 18:53:24 with our import of musl, comes an entirely new bootstrap method 2013-06-10 18:53:30 apk add musl-gcc for more details 2013-06-10 18:53:43 basically, we can build apks linked against musl from 2.7 2013-06-10 18:54:09 these can be used to bootstrap on archs we already support natively. 2013-06-10 18:54:28 on new archs, we can bootstrap from whatever. debian, for example. 2013-06-10 18:54:35 using the same method to yield a stage 1 2013-06-10 18:54:50 which is the same as old stage1, just a set of apks to use build-base packages :) 2013-06-11 06:05:35 apk add --> 100% cpu since upgrading 2013-06-11 06:05:54 its a non-trivial regression in apk. fabled is working on it 2013-06-11 06:06:02 okay, just wanted to be sure 2013-06-11 06:06:43 as i understand, it happens when there are many different possible resolve solutions 2013-06-11 06:06:47 but none are optimal 2013-06-11 06:07:43 likely you have some dependency conflict in your repos 2013-06-11 06:07:55 and some ov the pkgs in your 'world' depend on them 2013-06-11 06:08:44 this is incomplete, but it means we can bootstrap with aports. 2013-06-11 06:08:45 wow! 2013-06-11 06:08:49 nice work! 2013-06-11 06:08:58 and yes, better than crosstols! 2013-06-11 06:14:56 i'm going to bed now, i'll be back in a about 7 hours 2013-06-11 07:01:02 files from v2.4.11-40-g6da42db uploaded 2013-06-11 07:24:20 argh 2013-06-11 07:24:22 >>> mongodb*: Package size: 104.3 MB 2013-06-11 07:28:44 i remember i had that with mysql when i didnt subpkg the test suite 2013-06-11 07:39:10 this is only the /usr/bin/mongo* tools 2013-06-11 07:39:20 i havent seen anything bloated like this 2013-06-11 07:39:25 not really surprised though 2013-06-11 07:39:29 its c++ and javascript 2013-06-11 07:39:33 boost 2013-06-11 07:39:51 its easy to make c++ bloated 2013-06-11 07:40:28 and i think they link the same (bloated) object files in each tool statically 2013-06-11 07:40:33 so it adds up 2013-06-11 07:40:41 or multiplies to be exact 2013-06-11 07:47:47 ncopa: we don't have any newapkbuild helper for ruby gems right? 2013-06-11 07:48:01 nope 2013-06-11 07:48:24 i have just copied an existing and change the pkgname and version 2013-06-11 07:48:55 yeah, ill do the same. 2013-06-11 08:44:57 question: how should i go ahead if the sourcecode for some software is not available via download? 2013-06-11 08:45:28 put it into the a folder beside APKBUILD? 2013-06-11 08:45:55 i mean, there might be no fixed url where the sources could be downloaded 2013-06-11 08:49:46 StarWarsFan: you'll have to make your own copy on some fixed url 2013-06-11 08:50:06 we normally put it on http://dev.alpinelinux.org/archive 2013-06-11 08:50:24 ah i c 2013-06-11 08:51:15 so abuild will fetch it from there, right? 2013-06-11 08:51:23 yes 2013-06-11 09:34:52 files from v2.4.11-40-g6da42db uploaded 2013-06-11 11:28:33 ncopa, mongodb fails 2013-06-11 11:28:44 Checking for C library libsnappy... no 2013-06-11 11:28:44 can't find snappy library 2013-06-11 11:28:48 i noticed 2013-06-11 11:28:50 im on it 2013-06-11 11:28:52 k 2013-06-11 11:29:03 just wanted to be sure that you noticed it 2013-06-11 13:29:31 hey guys 2013-06-11 13:29:36 where is rc-service executable located? 2013-06-11 13:30:12 aha 2013-06-11 14:15:43 ncopa: do you mind if i add abyss-server to xmlrpc-c? 2013-06-11 14:37:18 if its not ok please revert it. 2013-06-11 14:47:47 as a subpkg... perfect! 2013-06-11 14:47:48 thanks! 2013-06-11 14:50:06 hacking on OpenNebula 2013-06-11 14:50:56 wonder if its worth it... 2013-06-11 15:00:28 isn't opennebula a giant java application that is ironically using alpine? 2013-06-11 15:06:54 i didnt see any java in the deps. 2013-06-11 15:07:02 only bindings 2013-06-11 15:07:20 @ncopa | >>> mongodb*: Package size: 104.3 MB 2013-06-11 15:07:27 WEBSCALE! 2013-06-11 15:10:06 https://jira.mongodb.org/browse/PYTHON-532 2013-06-11 15:11:03 do we really want to ship mongodb? :P 2013-06-11 15:22:55 ncopa: so, i should revert to an older apk-tools? 2013-06-11 15:30:28 ACTION wonders if he should push his half finished opennebula apkbuild. 2013-06-11 15:32:13 probable apk-tools regression target is commit cb98b55b7e25f45ed847bc32f3267d0c8fb033e5 2013-06-11 15:32:25 i think i have a fix 2013-06-11 15:38:18 libc: ASSIGN penalty too big: {0/6/0,25}+{0/0/0,12}>={0/5/0,16} 2013-06-11 15:38:20 BAM 2013-06-11 15:46:54 it's definitely this minimum penalty stuff, but my fix doesn't work :/ 2013-06-11 15:48:28 imo we should push back 2.3.4 for now? 2013-06-11 16:58:29 Hello...as I test Alpine-Linux Xen, I find that I can't do direct passthrough of the video card using xl pci-assignable-add. The system locks up... 2013-06-11 16:58:55 If I downgrade to 2.5 using apk upgrade -a ... all is good again. 2013-06-11 16:59:54 dmesg doesn't make anything jump out at me. I can pass through USB using pci-assignable-add, but not video. 2013-06-11 17:00:50 The question is... how can I downgrade one package at a time? 2013-06-11 17:01:10 I guess I could upgrade all, and hold a suspected package back using apk add 'asterisk<1.6.1', eh? 2013-06-11 17:02:05 BTW, upgrading from 2.5.4 to 2.6.0 solves abuild -r when trying to add sdl-dev. Works like a champ. 2013-06-11 17:02:45 ...but I was hoping to get both with the same package (pci passthrough AND sdl) :-) 2013-06-11 17:25:23 clandmeter, if you are testing ruby/thin, just to let know I have tested redmine successfully on unicorn, thin and puma(not yet in alpine) 2013-06-11 17:26:26 puma homepage has some benchmark showing it as least memory usage 2013-06-11 21:09:41 ncopa: im having trouble with bb getopt. I replaced it with util-linux version but its still using bb version. 2013-06-11 21:29:35 on the buildserver i mean. 2013-06-11 23:07:07 alpine linux support added to libreswan weeee 2013-06-11 23:07:12 now to wait for them to make a release 2013-06-12 05:40:38 moinmoin 2013-06-12 05:40:42 question: 2013-06-12 05:40:52 how is a *-doc package created? 2013-06-12 05:41:13 i've setup subpackage="$pkgname-doc" 2013-06-12 05:42:45 but build failed with message missing den directory $foo-doc 2013-06-12 05:42:53 so i created the directory on package() 2013-06-12 05:43:17 mkdir "$pkgdir-doc" 2013-06-12 05:43:26 now the -doc package ist created 2013-06-12 05:43:30 but it is empty 2013-06-12 05:43:56 i found nothing special on some other packages which create -doc subpackages 2013-06-12 05:44:00 so how to go ahead? 2013-06-12 05:44:18 how to "move" the man files to the -doc package? 2013-06-12 05:57:03 ok, install to "$pkgdir-doc/usr/share/man/man8/..." seems to solve the question... 2013-06-12 06:59:10 morning 2013-06-12 06:59:11 good morning 2013-06-12 06:59:19 how is a *-doc package created? 2013-06-12 06:59:22 hmm that was strange 2013-06-12 06:59:24 :) 2013-06-12 06:59:43 there is a package splitting function in abuild, named default_doc 2013-06-12 07:00:15 which moves common places like /usr/share/doc /usr/share/man from "$pkgdir" to "$subpkgdir" 2013-06-12 07:00:30 ah ok 2013-06-12 07:00:50 the package splitting is not very intuitive 2013-06-12 07:01:02 but basically 2013-06-12 07:02:40 ok 2013-06-12 07:02:50 if you add subpackages="$pkgname-foo" 2013-06-12 07:03:03 abuild will try execute "foo" as a split function 2013-06-12 07:03:09 so if the man pages where located on a non-common place, i'd have to install them the usual way 2013-06-12 07:03:34 i think you better beat the build system to place the man pages where they belong :) 2013-06-12 07:04:06 i don't want to touch a delivered makefile 2013-06-12 07:04:14 http://wiki.alpinelinux.org/wiki/APKBUILD_Reference#subpackages 2013-06-12 07:04:39 ok, great 2013-06-12 07:04:43 will try this 2013-06-12 07:04:45 thx! 2013-06-12 07:04:57 you can create a custom doc() {} function 2013-06-12 07:05:18 if you do: doc() { default_doc; # your stuff here } 2013-06-12 07:05:53 sometimes i patch makesfiles to place doc in proper location 2013-06-12 07:06:11 other times it is possible to do things like: make install MANDIR=/usr/share/man 2013-06-12 07:06:16 override things in makefile 2013-06-12 07:07:06 have to figure this out... 2013-06-12 07:19:47 ncopa: saw my issue? 2013-06-12 07:21:28 the usb kbd in emergency shell? 2013-06-12 07:22:18 in irc last night 2013-06-12 07:24:01 getopt? 2013-06-12 07:24:19 its a shell internal 2013-06-12 07:24:32 replace your shell with bash 2013-06-12 07:24:56 change first line from #!/bin/sh to #!/bin/bash 2013-06-12 07:24:58 and it should work 2013-06-12 07:25:06 which getopt 2013-06-12 07:25:07 /bin/getopt 2013-06-12 07:25:36 this getopt does not support --version 2013-06-12 07:25:41 hm 2013-06-12 07:25:44 you are right 2013-06-12 07:25:59 so i install util-linux 2013-06-12 07:26:00 $ which getopt 2013-06-12 07:26:10 /usr/bin/getopt 2013-06-12 07:26:11 which works local 2013-06-12 07:26:17 but not on build server 2013-06-12 07:32:25 build-edge:~$ which getopt 2013-06-12 07:32:25 build-edge:~$ getopt --version 2013-06-12 07:32:25 getopt from util-linux 2.22.2 2013-06-12 07:32:54 seems to work on build-edge? 2013-06-12 07:32:58 which build server has issues? 2013-06-12 07:36:30 afaik build-edge 2013-06-12 07:38:23 i patched the script to make it work 2013-06-12 07:39:22 looking at the patch 2013-06-12 07:39:46 looks like the original code was bad 2013-06-12 07:41:24 it works local when i install util-linux. but i agree the script aint pretty. 2013-06-12 07:44:28 i dunno 2013-06-12 09:49:40 i wonder if it would be possible to list all kernel modules for NICs that has support for PXE 2013-06-12 11:59:00 i'm thinking to remove calibre 2013-06-12 11:59:15 it's an old package, big, bloat, hard-to-maintain 2013-06-12 11:59:29 probably nobody uses it 2013-06-12 11:59:59 also has update every few weeks that breaks compilation 2013-06-12 12:00:10 it's quite frustrating :) 2013-06-12 12:20:38 fcolista: feel free to purge it 2013-06-12 12:24:16 or move it to unstable 2013-06-12 12:26:13 should rename unstable to unmaintained 2013-06-12 12:27:42 yup, good idea 2013-06-12 12:27:49 or unsupported 2013-06-12 12:28:41 unmaintained is more inviting 2013-06-12 12:28:51 agree 2013-06-12 12:28:55 like, if someone wants to maintain it, they could pick it up 2013-06-12 12:37:26 thx ncopa 2013-06-12 13:43:12 ncopa: is it allowed to create ruby-rack14? 2013-06-12 13:44:43 thats older rack, right? 2013-06-12 13:45:04 i suppose thats ok 2013-06-12 13:45:12 yes it is 2013-06-12 13:45:25 seems they can happily coexist 2013-06-12 13:45:54 then np 2013-06-12 14:07:21 barthalion, checked RT and found few companies using with uclibc successfully 2013-06-12 14:08:53 sounds doable and useful 2013-06-12 14:11:58 ncopa: can i upgrade ruby-json? 2013-06-12 14:12:13 i think so 2013-06-12 14:12:26 at least in edge 2013-06-12 14:12:43 clandmeter, rechecked puma and thin, seems they both use same parser (I missed this earlier) 2013-06-12 14:12:49 yes 2013-06-12 14:13:24 vkrishn: thin is a dep of opennebula 2013-06-12 14:13:35 and puma does not have daemon mode , I think 2013-06-12 14:14:03 I quite like it and planning to use with redmine 2013-06-12 14:14:10 but it is possible to switch web server (but frist need to get this thing going) 2013-06-12 14:14:46 ok 2013-06-12 14:15:20 vkrishn: whats wrong with unicorn? 2013-06-12 14:15:37 for personal use thin is much better 2013-06-12 14:15:47 unicorn is not suited for slow net connection 2013-06-12 14:16:01 define *much* 2013-06-12 14:18:07 for unicorn to work on high traffic/user add proxy like nginx in front of it 2013-06-12 14:18:24 I mean work effectively. 2013-06-12 14:18:31 read somewhere :) 2013-06-12 14:19:06 `dict much` 2013-06-12 14:23:16 ACTION moves much to /dev/null 2013-06-12 14:24:02 vkrishn: if you are the only one to use redmine on your lan, i would agree. 2013-06-12 15:26:57 hi 2013-06-12 15:45:07 oops 2013-06-12 15:45:09 wtf 2013-06-12 15:46:06 git did not do what i had in mind there. 2013-06-12 15:47:24 there 2013-06-12 15:47:34 stupid git 2013-06-12 20:13:45 files from v2.5.4-145-g3cbd2f2 uploaded 2013-06-12 20:19:41 algitbot: build 2.6-stable 2013-06-12 20:22:25 how can the checksum fail on 2.6 but be ok on 2.5 and edge?!? 2013-06-12 20:22:50 :) 2013-06-12 20:23:20 tdtrask: ncopa mentioned some cgit weirdness a while back 2013-06-12 20:23:23 maybe that? 2013-06-12 20:39:33 ACTION will leave it for ncopa :) 2013-06-12 20:40:13 anybody know how to fix current apk madness? 2013-06-12 20:40:27 kaniini: did you fix apk-tools? 2013-06-12 20:49:46 downgrading seems to work 2013-06-12 21:28:25 i have not 2013-06-12 21:29:09 it's introduced by minimum penalty stuff though 2013-06-12 21:29:14 i bisected 2013-06-12 21:38:23 nl infra is going down for a min. 2013-06-13 05:45:28 morning 2013-06-13 05:56:11 hi 2013-06-13 05:56:17 i get internal error on bugs.a.o 2013-06-13 05:58:34 i will reboot the bugs.a.o and www.a.o 2013-06-13 07:28:42 ncopa: i rebooted it yesterday, but vserver didnt come alive 2013-06-13 07:30:17 i rebooted the vserver now 2013-06-13 07:30:21 and it appears to have come up 2013-06-13 07:30:41 i upgraded it though 2013-06-13 07:32:02 fabled: i wonder what route to take for an apk-tools gui (gtk) 2013-06-13 07:32:18 either a gui could fork/exec apk 2013-06-13 07:32:40 or we could link to libapk.so 2013-06-13 07:32:49 we could use lua + lgi 2013-06-13 07:32:51 or vala 2013-06-13 07:33:02 or make a packagekit backend for apk 2013-06-13 07:34:20 the last sounds good 2013-06-13 07:34:29 i mean, make an apk backend for packagekit 2013-06-13 07:34:35 yeah, I know 2013-06-13 07:34:37 :) 2013-06-13 07:34:41 I don't know how much work would it require though 2013-06-13 07:34:53 i dont think that would be too difficult 2013-06-13 07:35:09 could unmask lots of features 2013-06-13 07:35:31 However designing and creating own frontend sounds like overengineering and waste of time 2013-06-13 07:35:44 otoh, i think apk tools has nice features that could not be presented easy in packagekit 2013-06-13 07:35:51 yea.. 2013-06-13 07:36:10 would be nice to have a tab with the 'world' 2013-06-13 07:36:20 and somethi like apk version output 2013-06-13 07:36:30 and a nice progressbar when apk upgrade 2013-06-13 07:36:32 And the main question is… do we really need gui? :p 2013-06-13 07:36:37 no :) 2013-06-13 07:36:47 i just miss it sometimes :) 2013-06-13 07:36:52 ah :D 2013-06-13 07:37:05 what i miss is a gui progressline :) 2013-06-13 07:37:38 i am fantisising about a gui installer 2013-06-13 07:37:43 someething simple 2013-06-13 07:38:42 i suppose using acf2 as backend for that would be nice 2013-06-13 07:39:02 then we could probably also have a curses installer 2013-06-13 07:39:10 ala yast 2013-06-13 07:39:21 well, its low prio 2013-06-13 07:39:49 acf2 has curses support? nice 2013-06-13 07:39:55 no 2013-06-13 07:39:58 not yet 2013-06-13 07:40:09 currently it is only an http rest api 2013-06-13 07:40:24 you coudl theoretically plug anything in front of it 2013-06-13 07:40:50 cool 2013-06-13 07:42:12 ncopa: apk is edge is broken 2013-06-13 07:42:24 it hangs? 2013-06-13 07:42:27 yes 2013-06-13 07:42:38 regression 2013-06-13 07:42:48 fabled is working on it as we speak 2013-06-13 07:42:57 kaniini said he found the issue by bisect 2013-06-13 07:42:58 call it a feature and rename apk to apkd 2013-06-13 07:43:06 lol 2013-06-13 07:43:34 clandmeter, kaniini: yes, there's a commit you can blame. while it makes it hang, it fixes another cases. 2013-06-13 07:43:41 i'm rewriting the whole solver.c part of apk 2013-06-13 07:44:13 will push out today my work, which is mostly in usable state with some caveats 2013-06-13 07:44:47 fabled: what was the other case it fixes? 2013-06-13 07:45:00 i wonder if we temp should revert that specific commit 2013-06-13 07:45:00 maybe its better to downgrade the one in edge? cause it completely stopped working for me. 2013-06-13 07:45:26 completely as ain "apk upgrade" 2013-06-13 07:45:34 -a 2013-06-13 07:45:44 ncopa: vserver is ok now? 2013-06-13 07:45:51 clandmeter: looks so yes 2013-06-13 07:46:04 kinda weird it didnt reboot yesterday 2013-06-13 07:46:14 clandmeter: i upgraded it before rebooting 2013-06-13 07:46:26 dunno what might have happened 2013-06-13 07:46:28 probably the one causing most breakage is http://git.alpinelinux.org/cgit/apk-tools/commit/?id=72cd34cf8148322837193c8e2b6da0cd96fe7788 2013-06-13 07:46:57 it was showing kvm pxe bios boot whatever 2013-06-13 07:47:05 clandmeter: do you mind if i purge crystalhd-git-grsec and enable crystalhd in linux-grsec? (its in staging) 2013-06-13 07:47:25 ncopa: sure 2013-06-13 07:47:31 i dont use it 2013-06-13 07:47:42 if immediate fix is needed, then it might be an idea the revert that in APKBUILD 2013-06-13 07:48:00 should have new version next week 2013-06-13 07:48:14 i think the same version is in 2.6 right? 2013-06-13 07:48:37 I had to pull the old apk from 2.6.0 iso 2013-06-13 07:48:57 fabled: looks like kaniini already pushed stuff to apk-tools git 2013-06-13 07:49:12 yeah, revert and revert of revert 2013-06-13 07:49:16 the musl fix looks useful though 2013-06-13 07:49:44 :) 2013-06-13 07:50:15 i think i might revert 72cd34cf8148322837193c8e2b6da0cd96fe7788 in apkbuild 2013-06-13 07:50:27 there was some case it fixed 2013-06-13 07:50:33 but i dont remember what 2013-06-13 07:52:03 yes 2013-06-13 07:53:27 if it was apk add ./local-1.0.apk 2013-06-13 07:53:36 then i'd revert it right away 2013-06-13 07:58:50 no it was something different... but might be an idea to revert anyway 2013-06-13 07:59:04 if the hang happens that often, it's worse than giving error in few places 2013-06-13 08:20:40 ncopa: have you tried to run ruby without gems installed? 2013-06-13 08:33:38 ncopa: i mean without ruby-gems installed. 2013-06-13 08:33:54 looks like ruby now depends on ruby-gems 2013-06-13 08:36:12 ugh 2013-06-13 08:36:17 looks so yes 2013-06-13 08:37:09 i guess its a matter of removing the subpkg right? 2013-06-13 08:37:52 and update all pkg's who have it in deps 2013-06-13 08:38:45 or make it a dummy pkg? 2013-06-13 08:39:08 or aadd it to depends? 2013-06-13 08:39:23 ruby-gems appears to depend on ruby-io-console 2013-06-13 08:40:25 is there a reason to subpkg it? 2013-06-13 08:40:47 the intention was to make it smaller 2013-06-13 08:41:18 it did :) 2013-06-13 08:41:35 but if its useless...? 2013-06-13 08:41:42 we could also add ruby-gems as a dep 2013-06-13 08:42:13 nah 2013-06-13 08:42:33 imho removing it makes more sence, but its more work. 2013-06-13 08:42:34 i suggest we remove ruby-gems and ruby-io-console subpkgs 2013-06-13 08:42:37 yes 2013-06-13 08:46:12 ncopa: can you remove them in one time? 2013-06-13 08:46:28 i think so 2013-06-13 08:46:30 oh 2013-06-13 08:46:35 should probably add: 2013-06-13 08:46:59 replaces="ruby-gems ruby-io-console" 2013-06-13 08:47:13 so upgrades works 2013-06-13 11:47:03 i wonder if i should enable ncurses in lua 2013-06-13 12:59:11 i am upgradeing lua to 5.2 2013-06-13 14:26:16 yay 2013-06-13 14:26:38 so we're going to ship 5.2 as default? 2013-06-13 14:30:32 yes i think so 2013-06-13 14:30:44 i have enabled the compat full 2013-06-13 14:30:59 i think most things will just work 2013-06-13 15:25:15 ok 2013-06-13 15:25:17 more to follow 2013-06-13 15:25:31 but likely early next week 2013-06-13 15:25:46 at least the rewrite should be in pretty decent and usable 2013-06-13 15:25:57 there's only 2 of 50 test cases not passing 2013-06-13 15:26:02 and those are related to error printing 2013-06-13 15:28:27 there's few caveats in the new code that all should be non-critical 2013-06-13 17:57:53 fabled: thanks for your hard work! 2013-06-13 17:58:27 clandmeter, apk-tools has some pending other things too that i plan to implement next week 2013-06-13 17:58:39 but if you can test the apk-tools.git head, feed back would be appreciated 2013-06-14 06:58:25 ncopa: i guess we cannot change dev permissions (mdev) with abuild? 2013-06-14 06:58:52 abuild? 2013-06-14 06:58:57 why? 2013-06-14 06:59:17 /dev/kvm is root:root 2013-06-14 06:59:38 i need it to be root:kvm or so 2013-06-14 06:59:50 so i have to manually add mdev.conf? 2013-06-14 07:00:17 edit i mean 2013-06-14 07:01:15 hm 2013-06-14 07:01:25 what package needs it? 2013-06-14 07:01:57 i have played with the idea of /etc/mdev.d/*.conf files 2013-06-14 07:02:18 and a trigger that will: cat /etc/mdev.d/*.conf > /etc/mdev.conf 2013-06-14 07:02:37 that way packages could ship an /etc/mdev.d/foo.conf file 2013-06-14 07:08:57 opennebula needs access to /dev/kvm 2013-06-14 07:11:46 ncopa: I guess there are more tools which would like to have access to /dev/kvm without running as root. 2013-06-14 07:21:44 we should add it in the default mdev.conf then 2013-06-14 07:33:28 ok ill add it. 2013-06-14 07:34:11 ncopa: if I add a user without a password, its locked by default? if so is there a way to unlock it so it can login with ssh key? 2013-06-14 07:34:38 now i have to passwd the user to unlock it. 2013-06-14 09:11:08 i will upgrade the build server so git.a.o might be unavailable for some minutes 2013-06-14 09:24:10 ok 2013-06-14 09:24:12 rebooting it now 2013-06-14 10:48:27 ncopa: edge builders gone? 2013-06-14 11:05:34 yups 2013-06-14 11:05:44 well 2013-06-14 11:05:48 only here in channel 2013-06-14 11:06:10 the use the buildmsg-* system usin zeromq nowdays 2013-06-14 11:06:33 i just pushed but no msg from algitbot 2013-06-14 11:08:47 algitbot: build master 2013-06-14 11:09:31 hum 2013-06-14 11:09:34 something is not working 2013-06-14 11:10:26 algitbot: build master 2013-06-14 11:21:44 algitbot: build master 2013-06-14 11:26:55 algitbot: build master 2013-06-14 15:26:06 the link for the devel category in the wiki: http://wiki.alpinelinux.org/wiki/Category:Development gives me a database error 2013-06-14 17:35:32 https://github.com/prymeroot/aports upgraded fail2ban, sudo, libxrender, libxvmc. It'd be nice if you pull. 2013-06-14 17:55:47 hi 2013-06-14 17:57:54 How can I change the setup-disk? 2013-06-14 17:57:59 How can I change the setup-disk script? 2013-06-14 17:58:18 Is there a git repo for this stuff? 2013-06-14 18:40:16 How can I send patches for http://git.alpinelinux.org/cgit/fab/alpine-conf/tree/ ? 2013-06-14 18:42:52 Maybe I know now, how 2013-06-14 19:41:13 prymeroot: does it build? I'm not eager to fix compilation errors 2013-06-14 19:41:22 testing your changes either 2013-06-14 19:46:27 barthalion: Which package? Also SOA #1 – every package built in chroot with no problems. 2013-06-14 19:46:59 whatever soa means in english 2013-06-14 19:51:00 files from v2.2.3-232-gad39fe8 uploaded 2013-06-14 19:51:31 files from v2.4.11-40-g6da42db uploaded 2013-06-14 19:51:38 files from v2.5.4-146-g55a1b25 uploaded 2013-06-14 19:51:38 files from v2.5.4-146-g55a1b25 uploaded 2013-06-14 19:52:00 wut 2013-06-14 19:55:45 ah, there was no notification about changes in 2.2/2.4/2.5, nvm 2013-06-14 19:56:36 barthalion: Oh, I mean it works on my Edge in chroot env. 2013-06-14 20:24:50 ncopa, ping 2013-06-14 20:25:19 ncopa, i remember a year or so back you had a patch to fix nfs on uclibc, whats' the state of that? 2013-06-14 20:25:26 i haven't looked at it in ages 2013-06-14 20:25:39 blueness: to fix which error? 2013-06-14 20:25:52 or bug* 2013-06-14 20:26:06 shafire, to be honest i don't remember now, it was something missing in uclibc 2013-06-14 20:26:29 let me ask this way, do you guys need any special patches to get nfs working on alpine? 2013-06-15 08:56:22 ncopa: rhel7 will use mariadb 2013-06-15 08:58:04 ncopa: I guess it's good idea to move to it, but it will require manual intervention for users moving from mysql 2013-06-15 08:58:45 ncopa: mysql_upgrade and mysqlcheck need to be executed, both require password to root of databases 2013-06-15 13:42:23 please git pull git://git.alpinelinux.org/fab/aports 2013-06-15 13:42:53 there are a couple of updated packages incl. new deps 2013-06-15 13:46:23 fabian_a: I prefer spaces over tabs in APKBUILDs :p 2013-06-15 13:46:26 thanks 2013-06-15 13:47:03 fabian_a: you probably saw my apk-tools module for ansible -- any feedback? 2013-06-15 13:50:08 build-2-5: build master 2013-06-15 13:50:25 fabian_a: test your changes next time, I assumed you did it =( 2013-06-15 13:50:31 algitbot: build master 2013-06-15 13:50:54 wut 2013-06-15 13:51:12 ncopa: How I'm supposed to fix it? 2013-06-15 13:53:24 algitbot: build this shit 2013-06-15 13:53:29 lol 2013-06-15 13:53:36 algitbot: build master 2013-06-15 13:53:45 it will fail 2013-06-15 13:54:01 I'm forecaster 2013-06-15 20:20:34 barthalion: i prefer spaces too but almost all APKBUILD are using tabs. the changes in the ansible APKBUILD are just for consistency 2013-06-15 20:22:54 barthalion: i tried to push an APK module into the ansible upstream but it was denied because for the developer aren't enough user around for apk. 2013-06-15 20:23:13 freeradius needs special dictionary for talking to JunOS. 2013-06-15 20:23:15 sadness. 2013-06-15 20:24:46 fabian_a: yeah, same 2013-06-15 20:25:15 fabian_a: my module is listed now here http://www.ansibleworks.com/docs/contrib.html 2013-06-15 20:25:47 so I guess it's good idea to take care of it and try to push it to upstream before next release… 2013-06-15 20:27:38 barthalion: i would like to have the apk module in the ansible core. 2013-06-15 20:27:47 mine is here https://github.com/fabaff/alpine-ansible/blob/master/modules/apk 2013-06-15 20:28:17 basically it's just a modified version of the pacman module 2013-06-15 20:28:52 I really need to improve pacman module… 2013-06-15 20:29:07 it may work for apk, but sucks for Arch itself 2013-06-15 20:29:48 my module is working but nothing more 2013-06-15 20:29:55 just like mine :p 2013-06-15 20:30:01 and I don't understand rationale behind rejecting apk module 2013-06-15 20:30:16 well, https://github.com/ansible/ansible/pull/2338 2013-06-15 20:30:17 Arch isn't server distro, so I don't see the point of supporting it 2013-06-15 20:31:05 i wasn't in the mood to start a long argument with the ansible upstream 2013-06-15 20:37:29 a question: was there a change that now packages in main can depend on packages in testing? you moved py-markupsafe from main to testing. a while back ncopa told me that this should be prevented. it seems that i^m not up to speed 2013-06-15 20:38:14 maybe that's why builders are broken now… 2013-06-15 20:39:29 let's see 2013-06-15 20:39:56 ACTION thanks barthalion for everything 2013-06-15 20:40:10 it would make sense 2013-06-15 20:40:52 fabian_a: np 2013-06-15 20:40:57 good night and enjoy your weekend 2013-06-15 20:41:44 same to you :) 2013-06-15 20:41:58 algitbot: buuuuillldd iiit 2013-06-15 20:42:48 wee, new errors 2013-06-15 20:43:08 OPEN STAK 2013-06-15 20:43:12 i mean hi 2013-06-15 20:43:19 lol 2013-06-15 20:43:21 hi :D 2013-06-15 20:50:14 \o/ 2013-06-15 20:50:28 algitbot: brofist 2013-06-15 20:50:36 good night :) 2013-06-16 13:46:33 woo-hoo, glusterfs isn't portable 2013-06-16 14:14:30 meh 2013-06-16 14:14:33 algitbot: build master 2013-06-16 18:35:22 sigh 2013-06-16 18:36:14 I should have predicted this 2013-06-16 18:41:49 I doubt that anyone is using it 2013-06-16 19:38:06 Ok, found a patch 2013-06-16 20:17:02 wee 2013-06-17 05:49:30 morning 2013-06-17 05:49:57 morning 2013-06-17 06:24:28 morning 2013-06-17 06:24:35 bugs.a.o gives internal errors 2013-06-17 10:01:21 ncopa: redmine logs show db errors 2013-06-17 10:12:37 ncopa: i ran the update process of redmine again (you told me last time you did). 2013-06-17 10:13:18 ncopa: it doesnt show errors anymore, guess some schema got updated on core or plugins. 2013-06-17 12:54:27 i think we will have to support both lua 5.1 and 5.2 for a while :-( 2013-06-17 14:01:04 ncopa: how do i get the pkgname from a apk file? 2013-06-17 14:01:40 ? 2013-06-17 14:01:50 if i have a package on disk 2013-06-17 14:02:01 can i use apk tools to tell me the pkgname? 2013-06-17 14:02:21 apk info --origin maybe? 2013-06-17 14:02:40 there shoud be a filed in there named 'origin' 2013-06-17 14:03:13 tar -O -zxf $apkfile .PKGINFO | grep origin 2013-06-17 14:03:41 im trying to prevent unpacking all apk's 2013-06-17 14:05:11 ncopa-desktop:~$ apk search --origin --exact lua-dev 2013-06-17 14:05:11 lua-5.1.5-r0 2013-06-18 12:32:32 ncopa: Are packages under unmaintained build by buildbots? 2013-06-18 12:32:56 barthalion: should not 2013-06-18 12:32:59 no they are not 2013-06-18 12:33:26 but my abump-kmods script detects it ... 2013-06-18 12:35:52 ah 2013-06-18 12:36:11 yeah, that's why I asked, I've seen the last commit 2013-06-18 12:38:51 I have a question about repository for stuff that is non-arch related... 2013-06-18 12:39:10 would it be a idea to have a repository for stuff that could be used for various alpine releases? 2013-06-18 12:39:18 That could save some 'backporting' 2013-06-18 12:40:16 how would you implement that? 2013-06-18 12:40:17 Hmm... thinking about it again... maybe it causes some trouble after all... 2013-06-18 12:40:50 unless it is packages that never changes (never updates) then yes 2013-06-18 12:40:52 Adding a row to /etc/apk/repositories that points to a place where there are apk's that are not realted to a specific alpine release 2013-06-18 12:41:06 so it has to be noarch 2013-06-18 12:41:12 and you cannot put luascripts there 2013-06-18 12:41:17 nor python 2013-06-18 12:41:25 so it is very limited what you can put there 2013-06-18 12:41:44 actually I think that pinning is enough 2013-06-18 12:41:45 In this specific case I think it should work just fine without lua/python/other 2013-06-18 12:42:51 But... still... maybe it will cause some problems after all for older releases. 2013-06-18 12:43:25 e.g. could acf-provisioning be programmed to handle some specific polycom firmwares (and not handling the newer firmwares) 2013-06-18 12:43:51 so... maybe the idea with a non-version related repository isn't a good idea afterall 2013-06-18 12:44:28 i wonder if its worth it 2013-06-18 12:44:55 because even if we put those in another subdir 2013-06-18 12:45:03 it'll get branched anyways 2013-06-18 12:47:57 Yeah... and now I found some other troubles with the idea too - so I don't think it's doable in this particular case either 2013-06-18 13:18:28 Question: What is the best way of doing a .patch file? 2013-06-18 13:18:46 Maybe it sounds like a strange question 2013-06-18 13:19:05 but what I am referring to is the first lines in the .patch file - which saids which files are to be modified 2013-06-18 13:19:19 in some .patch files the 'header' saids 2013-06-18 13:19:50 --- a/somfile 2013-06-18 13:19:50 +++ b/somfile 2013-06-18 13:20:15 And in some cases (if I remember it correctly) is that some say 2013-06-18 13:20:20 --- anotherfile 2013-06-18 13:20:23 +++ anotherfile 2013-06-18 13:20:31 (with no difference in the filenames) 2013-06-18 13:20:34 ah, it doesn't matter in fact 2013-06-18 13:20:46 create it using diff or git/svn/etc 2013-06-18 13:21:15 it matters the option "-pX" when applying the patch 2013-06-18 13:21:17 But he would still need to figure out which file he should patch! 2013-06-18 13:21:48 and it does 2013-06-18 13:21:48 "X" strips level X of directories from path 2013-06-18 13:21:51 so at least one of them needs to point to the right file to be patched (and I'm not sure if it's the first or second) 2013-06-18 13:22:22 it really depends on the location, where diff has been called 2013-06-18 13:26:36 rnalrd: Ok. Now I understood the -pX part 2013-06-18 13:26:51 looked at my APKBUILD and saw that it does 'patch -p1' 2013-06-18 13:27:12 so this means that my patch file could look like this: 2013-06-18 13:27:15 so with first example strips "a" and "b" dir levels 2013-06-18 13:27:31 --- org/myfile 2013-06-18 13:27:31 +++ myfile 2013-06-18 13:28:10 also second like should have "org" as prefix afaik 2013-06-18 13:28:16 second line* 2013-06-18 13:28:22 It works (I tested) 2013-06-18 13:28:28 k 2013-06-18 13:28:43 It's just that I was trying to figure out where to copy stuff before starting to modify anything 2013-06-18 13:29:39 so I did 'mkdir org && cp * org/ && vim myfile && diff org/myfile myfile > myfile.patch' 2013-06-18 13:29:43 and it worked great! 2013-06-18 13:29:45 Thanks! 2013-06-18 13:29:46 you could also do "git init ." 2013-06-18 13:29:54 change staff 2013-06-18 13:29:54 :-) 2013-06-18 13:30:02 git diff > mypatch 2013-06-18 13:30:11 rm -rf .git 2013-06-18 13:30:43 Hmm... good idea. It actually already is in git (working with aports tree) 2013-06-18 13:30:58 the downside is that this file is not in git repo yet (it's a new file) 2013-06-18 13:31:14 But I keep that in mind - could come in handy in other cases 2013-06-18 13:34:55 hm, there was a metapackage with gcc and the rest 2013-06-18 13:34:59 not alpine-sdk 2013-06-18 13:35:05 how is it called? 2013-06-18 13:36:05 build-base, nvm 2013-06-18 13:59:11 files from v2.5.4-154-g9d9e41e uploaded 2013-06-18 14:01:51 files from v2.5.4-154-g9d9e41e uploaded 2013-06-18 14:52:31 was reading about 'boot to qt5', can something like be included in al ? 2013-06-18 14:52:51 maybe without java ;) 2013-06-18 14:53:08 i have a qt5 pkg working i think 2013-06-18 14:53:13 or close to be working 2013-06-18 14:53:30 and samba-4 2013-06-18 14:54:00 cool, the light weight of UI fits in al overall design 2013-06-18 14:54:58 vkrishn: I pushed microcode packages to non-free 2013-06-18 14:55:22 vkrishn: and I'm working on linux-rt, I'll push it to testing during this week 2013-06-18 14:56:04 barthalion, sounds good 2013-06-18 14:56:20 would test once there 2013-06-18 14:56:56 I have ready APKBUILD, I need to set some things in kernel config 2013-06-18 14:58:01 ncopa: I'm wondering if it's possible to replace busybox modprobe with kmod 2013-06-18 14:58:10 probably 2013-06-18 14:58:44 btw 2013-06-18 14:58:51 boot to qt5 2013-06-18 14:59:03 i have been playing with the idea of 'boot to Lua' 2013-06-18 14:59:13 :) 2013-06-18 14:59:16 make a lib that has all syscalls 2013-06-18 14:59:25 and thats it 2013-06-18 14:59:58 then all userspace would be Lua 2013-06-18 15:00:37 i have also been dreaiming of an udev replacement impleneted in Lua 2013-06-18 15:00:47 the event handlers in lua 2013-06-18 15:00:50 maybe we try and test and see any new ideas get to it 2013-06-18 15:01:01 tiny C part that connects to netlink socket 2013-06-18 15:01:08 then forwards the rest to Lua 2013-06-18 15:01:53 then some ideas for parallelism 2013-06-18 15:01:57 since al has strong lua base, this could be something trying 2013-06-18 15:02:16 the hotplugger should only spawn N threads 2013-06-18 15:02:21 where N is the number of cores 2013-06-18 15:02:49 so you should never end up running more things in paralell than you have CPU cores 2013-06-18 15:04:08 maybe dub it as alpine-spin-offs 2013-06-18 15:04:40 unfortunally 2013-06-18 15:04:45 i have more ideas than time :) 2013-06-18 15:04:48 gotta go 2013-06-18 15:04:49 see you 2013-06-18 15:04:54 :) 2013-06-18 15:08:47 hmm... that almost made me planned to learn lua 2013-06-18 20:10:55 files from v2.5.4-155-gf8e24d2 uploaded 2013-06-18 20:11:05 files from v2.5.4-155-gf8e24d2 uploaded 2013-06-19 08:14:23 ncopa: we dont have ipmitool? or is there an alternative? 2013-06-19 09:31:59 hi 2013-06-19 09:32:05 clandmeter: i dont know 2013-06-19 09:34:18 looks like we want ipmitool 2013-06-19 09:34:50 and/or: http://ipmiutil.sourceforge.net 2013-06-19 09:36:16 im adding ipmitool 2013-06-19 09:36:22 http://ipmiutil.sourceforge.net/docs/ipmisw-compare.htm 2013-06-19 09:40:29 seen it 2013-06-19 10:22:37 fabled: new major version nr for apk-tools? 2013-06-19 10:28:40 clandmeter, yes 2013-06-19 10:28:50 i still have few things to work on today/tomorrow 2013-06-19 10:28:58 hope to tag something tomorrow or by latest on Friday 2013-06-19 10:29:10 either 2.4.0_rc1, or possibly 2.4.0 if things look good tomorrow 2013-06-19 10:30:39 77 files changed, 2860 insertions(+), 2997 deletions(-) 2013-06-19 10:31:19 which is pretty awesome since there's also new features in it 2013-06-19 10:39:08 nice 2013-06-19 11:01:18 files from v2.5.4-156-g9974c43 uploaded 2013-06-19 11:01:30 files from v2.5.4-156-g9974c43 uploaded 2013-06-19 12:02:17 jay! 2013-06-19 12:03:03 apk fetch --simulate 'lua-*' 2013-06-19 12:03:44 wow 2013-06-19 12:03:47 very useful 2013-06-19 12:03:58 apk del '*-dev' 2013-06-19 12:04:58 https://dpaste.de/7JHNa/ 2013-06-19 12:05:11 this is *awesome* 2013-06-19 12:28:44 :) 2013-06-19 12:29:09 add is trickier 2013-06-19 12:29:16 need to think something for it too 2013-06-19 12:29:28 i am sceptic to apk add '*' 2013-06-19 12:30:00 yes 2013-06-19 12:30:05 i am ok with wilcards only supported for fetch and del 2013-06-19 12:30:06 and it handles deps 2013-06-19 12:30:18 apk add "foo-*>1" ? 2013-06-19 12:30:22 yups 2013-06-19 12:30:23 or apk add !foo* 2013-06-19 12:30:28 so it's tricky 2013-06-19 12:30:36 i'm ok with not supporting it 2013-06-19 12:30:44 also needed is apk add foo.apk, that needs additional logic 2013-06-19 12:30:50 its not that useful for app 2013-06-19 12:30:53 add* 2013-06-19 12:30:54 true 2013-06-19 12:31:00 well 2013-06-19 12:31:05 some might like "apk add foobar-*" 2013-06-19 12:31:13 but even then, i'm slightly sceptic 2013-06-19 12:31:41 apk add $(apk search --exact 'foobar-*') 2013-06-19 12:31:43 or similar 2013-06-19 12:31:54 yes, it can be worked around 2013-06-19 12:32:34 hum 2013-06-19 12:32:40 apk search -q --exact 'lua-*' | xargs apk add --simulate 2013-06-19 12:32:42 works 2013-06-19 12:32:59 similar worked for del too 2013-06-19 12:33:14 but with del, fetch, etc. it's at least relatively safe 2013-06-19 12:33:24 apk del '*' 2013-06-19 12:33:26 :) 2013-06-19 12:33:38 yes, you can still shoot your self in foot 2013-06-19 12:33:48 like "cat world | xargs apk del" 2013-06-19 12:34:00 yup 2013-06-19 12:34:14 ok 2013-06-19 12:34:18 i'll close the bug then 2013-06-19 12:34:21 wildcards to apk add sounds like it creates more problems that it solves 2013-06-19 12:34:23 i think so 2013-06-19 12:34:27 say that we are sceptic against add "*" 2013-06-19 12:39:48 i still have two more changes 2013-06-19 12:39:51 and issue #1482 2013-06-19 12:40:18 though, only one of the changes i want in before tagging _rc1 2013-06-19 12:40:36 the other + the issue can wait until _rc2 or release 2013-06-19 12:44:38 If I am uncertain if I modified a file, and I would like to know what the original file looks like... how can I do that (e.g. if I get a .apk-new file I could compare it). 2013-06-19 12:44:38 In this case I am looking at /etc/conf.d/sysklogd 2013-06-19 12:44:52 Can I trigger apk to create a .apk-new file for me? 2013-06-19 12:45:42 Or.. maybe I can "apk verify /etc/conf.d/sysklogd" 2013-06-19 12:46:51 Nope... 2013-06-19 12:47:03 But in this case I could check how it looks at http://git.alpinelinux.org/cgit/aports/tree/main/sysklogd/sysklogd.confd 2013-06-19 12:47:36 (otoh - that doesn't necessary tell me how the file should look in the Alpine Linux version I am using) 2013-06-19 14:03:55 mhavela: maybe apk fix sysklogd 2013-06-19 14:04:25 i wrote a long email about lua 5.2 upgrade 2013-06-19 14:04:43 i am not sure what is the best way to move forward 2013-06-19 14:28:59 ncopa: http://sprunge.us/WUSC?sh is this ok to get the lastest aport packages? 2013-06-19 14:30:58 i dont think that will work for subpkg's :( 2013-06-19 14:33:35 you could also tar -O -zxf APKINDEX.tar.gz APKINDEX and parse the output 2013-06-19 14:33:50 i think there is a timestamp in the index 2013-06-19 14:34:55 what you do will give you latest aports 2013-06-19 14:35:40 maybe apkindex is better 2013-06-19 14:36:17 should be not too tricky with awk 2013-06-19 14:36:23 I want to update the db every x-period, so i dont have to import apk contents everyt x-period 2013-06-19 14:36:52 i mean instead of all apk's all the time. 2013-06-19 14:37:13 hm 2013-06-19 14:37:14 directory listing of 4 repos is pretty huge. 2013-06-19 14:37:25 yup 2013-06-19 14:37:26 my php parser is damn slow 2013-06-19 14:37:31 :) 2013-06-19 14:38:09 its using nested loops to filter paths from tar output 2013-06-19 14:38:21 and some packages have gaziilion entries 2013-06-19 14:38:47 ah you are importing the file contents? 2013-06-19 14:39:35 yes 2013-06-19 14:39:49 well not yet, just preparing 2013-06-19 14:40:50 i created a flat file similar like apkindex format 2013-06-19 14:41:06 p:pkgname f:filename 2013-06-19 14:41:23 i want to put that file on the mirrors 2013-06-19 14:41:28 for each repo 2013-06-19 14:42:46 pkgs=$(for i in *.apk; do [ $i -ot $yourflatfile ] || echo $i; done) 2013-06-19 14:42:47 anybody can use that for whatever they need. I have a small parser for a.o. which in the end import the file content and link it to the pkg browser. 2013-06-19 14:43:03 generate_index $pkgs > $yourflatfile 2013-06-19 14:43:29 i suppose it should be done with the apk index 2013-06-19 14:44:50 like an extended index 2013-06-19 14:45:53 http://nl.alpinelinux.org/.apk_contents_main_x86.txt.gz 2013-06-19 14:47:01 apk uses a different format for files 2013-06-19 14:47:34 have a look at /lib/apk/db/installed 2013-06-19 14:47:39 its more compressed 2013-06-19 14:47:43 what do you mean? 2013-06-19 14:47:50 let me check 2013-06-19 14:48:11 F:usr/bin 2013-06-19 14:48:19 R:bzip2 2013-06-19 14:48:26 R:bzegrep 2013-06-19 14:48:36 R:bzfgrep 2013-06-19 14:48:42 R:bzcat 2013-06-19 14:48:48 R:bzless 2013-06-19 14:48:51 etc 2013-06-19 14:48:54 rather than 2013-06-19 14:49:03 f:usr/bin/bzip2 2013-06-19 14:49:13 f:usr/bin/bzegrep 2013-06-19 14:49:17 how does it now know the path? 2013-06-19 14:49:27 F: 2013-06-19 14:49:43 R: 2013-06-19 14:50:23 ok 2013-06-19 14:51:14 save some space if you have a long path to a dir with many files 2013-06-19 14:51:26 saves* 2013-06-19 14:51:43 yes, but needs more coding experiance.. :) 2013-06-19 14:52:52 i wonder if we should make a feature request to fabled 2013-06-19 14:53:05 he is working on apk now 2013-06-19 14:53:19 something like apk index --file-index 2013-06-19 14:53:55 with the same format as in db/installed? 2013-06-19 14:54:22 yes 2013-06-19 14:55:06 in any case 2013-06-19 14:55:28 you can compare all *.apk timestamp with the timestamp of your index 2013-06-19 14:55:55 everythign thats newer than last-modified-time of your index could be considered 2013-06-19 14:56:03 yes, well actually im doing that already for the package db on a.o 2013-06-19 14:56:14 so maybe i should just hook in there 2013-06-19 14:58:24 do you mind (for the time beeing) i will put those apk_contents files on our mirror? 2013-06-19 15:00:26 going home. ttyt. 2013-06-19 21:03:07 (10/36) Installing krb5-conf (1.0-r0) 2013-06-19 21:03:07 (11/36) Installing sqlite-libs (3.7.17-r0) 2013-06-19 21:03:07 (12/36) Installing heimdal-libs (1.5.2-r7) 2013-06-19 21:03:08 huh 2013-06-19 21:03:18 what were those depends added to o.O 2013-06-20 07:54:46 i think we should set up an alpine-users mailing list 2013-06-20 10:28:14 ... 2013-06-20 10:28:21 perl 5.18 2013-06-20 13:31:08 ok i think finally have a plan for lua 2013-06-20 13:31:25 lua 5.2 that is 2013-06-20 13:32:15 i htink i will rename 'lua' pkg to lua5.1 2013-06-20 13:32:34 then have a meta package named 'lua' which only has symlinks 2013-06-20 13:32:57 ncopa, it could be also lua-common that has a trigger and updates the symlinks 2013-06-20 13:32:57 the lua-dev meta pkg will only depend on lua5.1-dev 2013-06-20 13:33:10 ncopa, no meta pkg; use provides instead. 2013-06-20 13:33:38 (unless there's other reasons for the meta pkg) 2013-06-20 13:34:37 containing symlinks sounds like a reason 2013-06-20 13:34:52 does that mean both lua 5.1 and 5.2 for v2.7? 2013-06-20 13:35:03 vkrishn: dunno yet 2013-06-20 13:35:07 hopefully not 2013-06-20 13:35:09 why don't we have provides described on wiki? 2013-06-20 13:35:22 what is up-to-date reference then? 2013-06-20 13:35:40 fabled: i dont think provides works in fully automated systems 2013-06-20 13:35:44 (read build servers) 2013-06-20 13:35:49 why not? 2013-06-20 13:35:59 it will exit and ask user to install one? 2013-06-20 13:36:02 no 2013-06-20 13:36:19 if you specify provides=lua-dev=5.2 2013-06-20 13:36:33 as long as the provided name is specified with version, it gets automatically installed 2013-06-20 13:37:07 barthalion, http://git.alpinelinux.org/cgit/apk-tools 2013-06-20 13:37:09 ;) 2013-06-20 13:37:13 though yes, wiki needs updating 2013-06-20 13:37:17 I thought so :p 2013-06-20 13:37:19 so the lua5.1-dev pkg could have provides=lua-dev-5.1 2013-06-20 13:37:40 and lua5.2-dev pkg would have provides=lua-dev-5.2 2013-06-20 13:38:08 provides=lua-dev=5.1 and provides=lua-dev=5.2 2013-06-20 13:38:13 and if you do: apk add lua-dev 2013-06-20 13:38:22 seeing last few months updates for various major packages, the phase seems a disruptive in nature, I was thinking why not mark v2.7 as transition release (just thought) 2013-06-20 13:38:25 it'll pull in the latest? 2013-06-20 13:38:52 vkrishn: I suggested something similiar 2013-06-20 13:39:05 and add both packages like python 2.7x+3x... etc 2013-06-20 13:39:22 increase the release date by 3/4months 2013-06-20 13:39:23 i think we will end up with using v2.7 as transition for lua 2013-06-20 13:39:31 ncopa, yes, apk add lua-dev would pull in lua5.2-dev since the latest provided versions comes from it 2013-06-20 13:40:04 also if possible kmscon/wayland support 2013-06-20 13:40:06 and apk add 'lua' would pull in provides=lua=5.2 2013-06-20 13:40:10 also we should decide what we do with mysql, but I'll send RFC to alpine-devel 2013-06-20 13:40:12 which would break much 2013-06-20 13:40:13 ncopa, yes 2013-06-20 13:40:15 well 2013-06-20 13:40:22 depends how define them 2013-06-20 13:40:37 it'll break the build msg on build servers 2013-06-20 13:40:39 provides=foo=version makes it act as-if it was a real package with that 2013-06-20 13:40:55 it also means they cannot co-exist 2013-06-20 13:41:00 yes 2013-06-20 13:41:09 so it would likely work for lua-dev 2013-06-20 13:41:11 but not lua 2013-06-20 13:41:19 right 2013-06-20 13:41:31 so my though with meta package was 2013-06-20 13:41:42 we have lua-dev that has symlinks 2013-06-20 13:41:52 lua.pc -> lua5.1.pc 2013-06-20 13:42:04 and /usr/bin/lua -> lua5.1 2013-06-20 13:42:28 and most importantly, lua-dev must *no* depend on 'lua' 2013-06-20 13:42:38 which it does now 2013-06-20 13:42:39 hmm 2013-06-20 13:42:41 well 2013-06-20 13:42:44 it might do 2013-06-20 13:42:55 but then lua5.2 cannot provide /usr/bin/lua 2013-06-20 13:43:13 right 2013-06-20 13:43:22 sounds like you have it thought out then 2013-06-20 13:43:30 i'm getting there :) 2013-06-20 13:43:38 took me all week 2013-06-20 13:43:45 and yes, symlinks in the -dev make it a validly a package. and it's no longer a "meta" package then. 2013-06-20 13:43:54 yes 2013-06-20 13:44:29 it woudl contain the current default '#!/usr/bin/lua' 2013-06-20 13:45:10 once we have migrated all packages, or changed to lua5.1-dev in makedepends 2013-06-20 13:45:19 we can point the 'lua' to 5.2 2013-06-20 13:45:42 i think that is the only way 2013-06-20 13:45:49 only way without breaking stuff 2013-06-20 13:45:57 also 2013-06-20 13:46:09 its pretty nice to be able to have both 5.1 and 5.2 installed in parallel 2013-06-20 13:46:17 in case you develop a lua module 2013-06-20 13:46:25 you might want test that it works for both versions 2013-06-20 13:47:11 could maybe do something similar with python3 2013-06-20 13:47:11 true 2013-06-20 13:47:54 ok good 2013-06-20 13:48:13 and i'll have to redesign my lua5.2 apkbuild *again*... 2013-06-20 13:48:58 oh while are at it, lua- packages should not have depend=lua 2013-06-20 13:49:12 because they strictly does not depend on the interpreter 2013-06-20 15:18:10 clandmeter: can I log in forum without registering ? I am already registered on wiki/bugs.a.o 2013-06-20 15:18:25 using same user/pass 2013-06-20 15:19:02 vkrishn: not yet 2013-06-20 15:19:06 you will have to signup again. 2013-06-20 15:19:19 ok 2013-06-20 16:05:07 yea, my 1st post :) 2013-06-20 16:07:06 I just upgraded my usb-stick which used to hold AL 2.4.3 with a encrypted apkovl 2013-06-20 16:07:14 Upgraded it to 2.6.0_rc2 2013-06-20 16:07:30 (which might be a bad idea because that after all is a Release candidate) 2013-06-20 16:07:32 anyway... 2013-06-20 16:07:37 clandmeter: I keep getting "You must be a human, not a spam bot, to submit forms on this website. ...blabla.." 2013-06-20 16:07:52 cannot post :( 2013-06-20 16:08:06 when booting, I get asked for password to unlock the encrypted apkovl - but the keyboard stopped working (it's a USB-keyboard) 2013-06-20 16:08:15 So - I am not able to logon no more 2013-06-20 16:08:26 ACTION fetches a PS2 keyboard to verify if it works 2013-06-20 16:09:09 Yup - a PS2 keyboard did a difference 2013-06-20 16:15:32 clandmeter: ping 2013-06-20 18:58:27 smokeping 2.6.9 needs to be backported btw 2013-06-20 18:59:01 And I think there's something wrong with the init.d script for smokeping 2013-06-20 18:59:06 just don't remember what it was 2013-06-20 19:00:29 mhavela: no slave support 2013-06-20 19:00:42 i just launched my slaves by hand tho 2013-06-20 19:00:52 http://monitor.tortois.es/smokeping/smokeping.cgi?displaymode=n%3Cscript%3Ealert%28%27lol%27%29%3C/script%3E;start=2013-06-20%2013:02;end=now;target=SecuredServers.SecuredServersPhoenix 2013-06-20 19:00:56 at least this XSS is fixed now ;) 2013-06-20 21:13:12 kaniini: so why you won't backport it? :p 2013-06-20 21:22:00 barthalion: i'll do it in a bit. 2013-06-20 21:30:49 kaniini: Please have a extra eye on the init.d script too if you do any modifications to the APKBUILD (just in case I was right about there being something wrong with it) 2013-06-20 21:31:17 vkrishn: if you are reading backlog. Failed 1 of 5 recipes [obscure_url] from "default" recipe book. 2013-06-20 21:31:31 I /think/ it was something with the path to the config file 2013-06-20 22:41:57 clandmeter: re: obscure_url, error from drupal/forum? version of drupal ? 2013-06-21 06:15:07 morning 2013-06-21 11:56:24 vkrishn: let me know which content you tried to post. ill see what happend. 2013-06-21 11:58:08 algitbot: build master 2013-06-21 12:04:27 files from v2.5.4-157-g71014d6 uploaded 2013-06-21 12:06:50 files from v2.4.11-41-gb9073a5 uploaded 2013-06-21 12:07:08 files from v2.3.6-244-gd18986d uploaded 2013-06-21 12:07:40 files from v2.3.6-244-gd18986d uploaded 2013-06-21 12:15:54 hmpf 2013-06-21 12:15:59 xen security issues 2013-06-21 12:16:29 royger: do you know if the xsa patches should be applied to the hypervisor (eg xen package) or the kernel (eg linux-grsec)? 2013-06-21 12:16:44 xsa-55 2013-06-21 12:16:49 http://lists.xen.org/archives/html/xen-devel/2013-06/msg01626.html 2013-06-21 12:18:32 ok looks like xen 2013-06-21 12:36:52 ncopa: yes, that's Xen 2013-06-21 12:37:14 ncopa: most XSA are for Xen, unless otherwise specified in the advisory 2013-06-21 12:37:37 ok 2013-06-21 12:37:38 ncopa: sorry, I've been dragged into other stuff lately and haven't been able to push any security patches recently 2013-06-21 12:38:23 kocka help me to report them 2013-06-21 13:27:02 ncopa: is `apk cache sync` still need on new usb installs? 2013-06-21 13:27:14 i think so 2013-06-21 13:27:33 didnt you add some logic to setup-alpine? 2013-06-21 13:27:39 i remember i reported it some time ago. 2013-06-21 13:27:54 hm 2013-06-21 13:27:58 i dont remember 2013-06-21 13:28:01 check git log 2013-06-21 13:28:16 might be its no longer needed 2013-06-21 13:28:40 i wonder why the usbkbd kernels were not added to the iso 2013-06-21 13:28:44 initramfs 2013-06-21 13:28:46 i guess it still is: http://alpinelinux.org/forum/installation/installation-issues-alpine-xen-usb 2013-06-21 15:13:06 files from v2.4.11-45-g784f0f4 uploaded 2013-06-21 15:27:47 failed to build v2.4.11-45-g784f0f4. http://distfiles.alpinelinux.org/distfiles/build64-2-4.txt 2013-06-21 23:07:44 files from v2.4.11-46-g45ce05e uploaded 2013-06-21 23:08:18 files from v2.3.6-245-g7faa1fc uploaded 2013-06-21 23:08:19 files from v2.3.6-245-g7faa1fc uploaded 2013-06-21 23:53:09 failed to build v2.4.11-46-g45ce05e. http://distfiles.alpinelinux.org/distfiles/build64-2-4.txt 2013-06-22 00:01:59 hi 2013-06-22 00:02:07 just installed the latest 2.6 iso on a virtualbox machine 2013-06-22 00:02:14 and it crashes with an ooops 2013-06-22 00:06:04 ah, wait a minute 2013-06-22 00:11:44 nope 2013-06-22 00:11:49 it's not booting inside virtualbox 2013-06-22 00:12:51 http://i.imgur.com/7YAAwAC.png 2013-06-22 00:44:03 the problem are the dependency modules of xfs 2013-06-22 14:19:28 grimeton: hm 2013-06-22 14:19:36 grimeton: take a look at /etc/mkinitfs/mkinitfs.conf 2013-06-22 14:19:43 grimeton: you should add xfs there 2013-06-22 17:18:40 hi 2013-06-24 06:08:59 hi 2013-06-24 06:09:10 algitbot: build 2.6-stable 2013-06-24 06:10:11 algitbot: build 2.6-stable 2013-06-24 08:02:23 failed to build v2.4.11-46-g45ce05e. http://distfiles.alpinelinux.org/distfiles/build64-2-4.txt 2013-06-24 08:09:10 ncopa: did you take a look at pacemaker? 2013-06-24 08:09:44 not yet 2013-06-24 08:09:53 i see y ou upgraded drbd 2013-06-24 08:09:55 i'm on it now 2013-06-24 08:10:03 pacemaker has issues 2013-06-24 08:10:10 its inside libsomething 2013-06-24 08:10:26 libqc i think 2013-06-24 08:11:12 libqb 2013-06-24 08:12:53 build of pacemaker fails. the cmd it executes segaults in libqb afaik. 2013-06-24 08:17:50 cluster-glue needs update/fix too 2013-06-24 08:29:18 doesnt fix my issue 2013-06-24 12:10:18 clandmeter: re: forum, cannot create new posts in "General discussion", but able to post reply to existing one. 2013-06-24 12:10:33 hi 2013-06-24 12:10:36 hope the bugs been fixed 2013-06-24 12:10:39 hi 2013-06-24 12:10:55 can you sprunge your post content? 2013-06-24 12:13:07 https://dpaste.de/R4gMK/ 2013-06-24 12:22:52 any spam filter being applied? 2013-06-24 12:23:10 we use a module for spam 2013-06-24 12:23:17 botcha 2013-06-24 12:23:58 drupal module 2013-06-24 13:01:08 vkrishn: i disabled osbsure url check 2013-06-24 13:01:13 seems to be going wrong 2013-06-24 13:01:20 i posted your question in general 2013-06-24 13:01:47 thanks 2013-06-24 13:44:27 is there any official alpine BNC I can apply to ? ok to me even if it just allows to AL irc's 2013-06-24 13:45:31 I mean request to ? 2013-06-24 14:04:12 vkrishn: you mean BNC as a service? 2013-06-24 14:04:44 yes, like znc 2013-06-24 14:05:17 are there projects that provide services like that? 2013-06-24 14:06:31 not sure, but if limited to AL official irc's should be an abuse 2013-06-24 14:06:41 it better than using/viewing the logs and doing ghost appearance 2013-06-24 14:07:16 I mean for slow or limited internet connections 2013-06-24 14:07:46 show not be an abuse* 2013-06-24 14:08:20 and if set some limit to expiry, even better 2013-06-24 14:08:33 like 3months not logged... etc 2013-06-24 14:12:48 eg quassel - seems nice 2013-06-24 14:22:42 algitbot: build 2.6-stable 2013-06-24 14:30:27 ok now ghost is disappearing in logs check only mode ;) 2013-06-24 16:34:58 failed to build v2.4.11-46-g45ce05e. http://distfiles.alpinelinux.org/distfiles/build64-2-4.txt 2013-06-25 05:21:36 all of this is running on various alpine machines: http://turtle.dereferenced.org/~nenolod/panel2-20130624.ogv 2013-06-25 06:32:24 nice video kaniini 2013-06-25 07:52:38 failed to build v2.4.11-46-g45ce05e. http://distfiles.alpinelinux.org/distfiles/build64-2-4.txt 2013-06-25 09:09:06 kaniini: very cool video 2013-06-25 09:09:39 that is of course a complex java application :P 2013-06-25 09:09:41 erm 2013-06-25 09:09:43 not java 2013-06-25 09:09:45 bahahahahah 2013-06-25 09:09:48 python + uwsgi 2013-06-25 09:09:48 lol 2013-06-25 09:09:53 cool! 2013-06-25 09:09:58 sorry i am used to typing java after complex 2013-06-25 09:10:13 we're probably gonna use uwsgi for acf2 2013-06-25 09:10:14 with lua 2013-06-25 09:10:50 i patched uwsgi for lua 5.2 yesterday 2013-06-25 09:11:05 anyway, that's alpine, python, uwsgi, gevent/geventwebsocket and xen 2013-06-25 09:11:11 very nice 2013-06-25 09:11:44 going to start merging some of that stuff into alpine 2.6 (since it's all open source code anyway) 2013-06-25 09:11:59 oh, also the box i did the screencast on runs alpine 2013-06-25 09:12:01 haha 2013-06-25 09:12:04 :) 2013-06-25 09:12:08 alpine + gnome? 2013-06-25 09:12:23 man... youre crazy for real :) 2013-06-25 09:12:58 put it on youtube ;) 2013-06-25 09:14:29 it's already there 2013-06-25 09:15:14 where? 2013-06-25 09:15:43 found it 2013-06-25 09:54:25 failed to build v2.4.11-46-g45ce05e. http://distfiles.alpinelinux.org/distfiles/build64-2-4.txt 2013-06-25 12:50:15 failed to build v2.4.11-46-g45ce05e. http://distfiles.alpinelinux.org/distfiles/build64-2-4.txt 2013-06-25 13:27:36 kaniini: you are still working on the ARM port right? 2013-06-25 13:28:27 kaniini: I've ordered one of this boards http://cubieboard.org/2013/06/19/cubieboard2-is-here/ they are cheap, quite powerful and should be able to run Xen 2013-06-25 14:16:40 failed to build v2.4.11-46-g45ce05e. http://distfiles.alpinelinux.org/distfiles/build64-2-4.txt 2013-06-25 21:02:54 kaniini: you around ? 2013-06-25 21:04:04 saw the panel2 video, nice 2013-06-26 03:47:05 Has alpine been tested on a google cr-48? 2013-06-26 06:25:34 I wouldn't count on it 2013-06-26 06:31:21 In AL 2.4.11 (and maybe also other versions) then openrc will not know the status of shorewall if the user does e.g. 'shorewall stop' (the only way of having openrc know the status is to use the '/etc/init.d/shorewall' script) 2013-06-26 07:29:37 failed to build v2.4.11-46-g45ce05e. http://distfiles.alpinelinux.org/distfiles/build64-2-4.txt 2013-06-26 07:42:48 files from v2.4.11-47-g3015aac uploaded 2013-06-26 08:22:57 files from v2.4.11-47-g3015aac uploaded 2013-06-26 08:56:00 Heartmender: i have a cr-48 with a nail in it 2013-06-26 11:13:19 files from v2.4.11-51-gab94a52 uploaded 2013-06-26 11:16:55 files from v2.4.11-51-gab94a52 uploaded 2013-06-26 11:20:36 files from v2.5.4-163-gc78cd17 uploaded 2013-06-26 11:24:04 files from v2.5.4-163-gc78cd17 uploaded 2013-06-26 12:15:04 files from v2.4.11-52-g386d947 uploaded 2013-06-26 12:28:11 testing/uWSGI newer versions are available, both LTS, stable 2013-06-26 12:39:42 ncopa: do pinned packages get upgraded if higher version is available in other repos? 2013-06-26 12:42:42 im going to downgrade ruby-rack to 1.4.5 to satisfy redmine. anybody against it should raise his hands within 3 seconds. 2013-06-26 12:55:42 files from v2.3.6-248-gdac4485 uploaded 2013-06-26 12:56:11 nice, pinning sticks to its own repo. 2013-06-26 13:02:35 files from v2.3.6-248-gdac4485 uploaded 2013-06-26 18:21:47 is there also an alpine pxe image? 2013-06-26 18:47:23 Hello 2013-06-26 18:47:44 i have created two apkbuild files for erlang and yaws 2013-06-26 18:47:53 seems to be they work perferct 2013-06-26 18:48:57 how can i pull them into git ? 2013-06-26 18:49:57 i did everything that described here http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2013-06-26 18:50:37 but I did not undestand in which directory I must place new files: main, testing, unmaintained 2013-06-26 18:51:18 testing will be good place 2013-06-26 18:51:33 push it then and ask here for pull 2013-06-26 18:51:43 or just send your changes to alpine-devel 2013-06-26 18:52:04 can you explain how do this properly 2013-06-26 18:52:27 I put files in testing/erlang15 for example 2013-06-26 20:52:10 well, seems to be all fine :) 2013-06-26 21:12:48 good night ;-) 2013-06-27 09:07:02 how can i avoid framebuffer console? nofb and nomodeset won't works 2013-06-27 09:09:55 trying with F00 (80x25) 2013-06-27 10:02:05 hm 2013-06-27 10:02:11 i thought nomodeset should done it 2013-06-27 13:14:05 ok seems that the problem is not related to framebuffer. 2013-06-27 13:15:07 with nomodeset i can see the rc services starting, but when alpine clear the screen before the prompt, everything becomes black and the keyboard is not responding (caps lock and num lock won't works) 2013-06-27 13:15:19 alpine 2.6.1 x86 2013-06-27 13:15:22 from cd 2013-06-27 13:15:31 anyone else has experienced the same? 2013-06-27 13:25:17 files from v2.5.4-164-g7e59376 uploaded 2013-06-27 13:25:29 files from v2.5.4-164-g7e59376 uploaded 2013-06-27 13:29:50 files from v2.4.11-53-ge4a6a3b uploaded 2013-06-27 13:33:13 files from v2.3.6-250-g142cf74 uploaded 2013-06-27 13:33:21 files from v2.3.6-250-g142cf74 uploaded 2013-06-27 13:33:40 tdtrask: any plans to port ACF to lua 5.2? 2013-06-27 13:35:15 ncopa: not yet, no bandwidth 2013-06-27 13:35:31 ACTION isn't sure what would need to change 2013-06-27 13:38:21 its time to think in that direction 2013-06-27 13:38:33 yes, I'm aware :) 2013-06-27 13:38:42 any new code should be 5.2 compatible 2013-06-27 13:39:18 there are some pages that explains the difference 2013-06-27 13:39:35 i think major things are module() 2013-06-27 13:39:51 instead of using module(...,package.seall) 2013-06-27 13:40:07 the module should return a table of the public api 2013-06-27 13:40:20 return { func1, func2, ....} 2013-06-27 13:40:39 http://lua-users.org/wiki/ModuleDefinition 2013-06-27 13:41:45 and similar with require 2013-06-27 13:41:48 replace: 2013-06-27 13:41:55 require('mymod') 2013-06-27 13:41:57 with: 2013-06-27 13:42:05 mymod = require('mymod') 2013-06-27 13:42:18 but we have compat enabled for now 2013-06-27 13:42:29 so i think most things will just work 2013-06-27 13:42:52 i wonder if i should build a haserl5.2 package 2013-06-27 13:42:58 so you could test what happens 2013-06-27 15:27:14 ncopa: any links to those pages? 2013-06-28 05:56:01 tdtrask: http://lua-users.org/wiki/LuaFiveTwo 2013-06-28 05:56:16 http://www.lua.org/manual/5.2/manual.html#8 2013-06-28 11:21:03 kaniini, what's up with the arm build? 2013-06-28 11:21:38 i just got my RPi, and am thinking of making alpine tick on it 2013-06-28 11:21:58 though for my immediate needs using raspbian is also an option 2013-06-28 12:34:14 seems that we'd need to ship specific bootloader + binary blob for the videocore chip 2013-06-28 12:34:27 and have patched kernel (likely no grsec) 2013-06-28 12:34:56 but appears that we can build from source. the closed source stuff is in the videocore binary blob only. 2013-06-28 12:35:08 and the kernel drivers + userland is just shim to talk to that blob 2013-06-28 16:21:27 it isnt workink: svn export svn://svn.alpinelinux.org/acf/core/trunk/www/cgi-bin/mvc.lua 2013-06-28 16:21:28 is it? 2013-06-28 16:22:38 I'm following this: http://wiki.alpinelinux.org/wiki/ACF_mvc.lua_example 2013-06-28 16:22:41 trying 2013-06-28 16:43:56 alacerda: that's quite old, we haven't used SVN in a long time 2013-06-28 16:44:14 sorry about the out-of-date docs 2013-06-28 16:45:18 http://git.alpinelinux.org/cgit/acf-core/plain/lua/mvc.lua is the latest mvc.lua 2013-06-28 16:45:57 ACTION hasn't gone through those docs in a long time :( 2013-06-28 16:48:12 tdtrask: thaks... 2013-06-28 16:48:44 but, if i follow this doc, fixing this mvc.lia url, will it work? 2013-06-28 16:48:53 may i follow the doc yet? 2013-06-28 16:56:47 you can try :) 2013-06-28 16:56:58 hehe... ok 2013-06-28 16:57:14 alacerda: it will progressively help you learn, but it's not a recommended way of doing anything 2013-06-28 16:57:41 are you trying to create a new ACF? 2013-06-28 16:57:48 a new acf module 2013-06-28 16:58:16 for what package? 2013-06-28 16:59:12 here we'll need to develop a specific solution (to integrate FreeSWITCH with Siemens HiPath4000). And I wanna create a module for manage it through ACF 2013-06-28 17:00:03 "best" method is to find another ACF with similar functionality and start from there 2013-06-28 17:00:45 hum... you mean clone acf-freeswitch and just change things? 2013-06-28 17:00:57 acf-freeswitch is an example* 2013-06-28 17:01:41 ACTION isn't sure what "manage it" entails 2013-06-28 17:01:54 so, hard to recommend the best example 2013-06-28 17:02:11 do you think this way i'll understand MVC model and other stuffs? I'm starting from scratch.. i've never done any lua MVC aplication 2013-06-28 17:02:22 application* 2013-06-28 17:02:46 might be a good idea to read through the docs, because many concepts are the same 2013-06-28 17:03:27 and as next step, clone another acf-module and try to change things... Ok! I'll try. 2013-06-28 17:03:36 thanks for the direction 2013-06-28 17:04:12 "manage it" means some things like see my solutions' database records 2013-06-28 17:05:28 several ACFs deal with databases (mostly postgres) 2013-06-28 17:05:44 kamailio, provisioning, weblog, freeswitch-vmail, ... 2013-06-28 17:06:33 hum rum... i'll see how they do that 2013-06-28 17:06:48 thanks 2013-06-28 17:08:11 welcome 2013-06-28 17:08:22 feel free to ping me if you run into trouble 2013-06-28 17:08:36 ; -) 2013-06-28 22:01:55 https://github.com/prymeroot/aports upgraded qemu to 1.5.1, pull if you want 2013-06-28 22:34:40 imagemagick was upgraded too 2013-06-29 08:05:31 another updates: json-c (0.11) and rsyslog (7.4.1) 2013-06-29 09:56:12 and another: freetype (2.5.0.1), logrotate (3.8.5), varnish (3.0.4). Yep, lots of incredibly important updates :p 2013-06-29 09:56:17 2013-06-29 14:59:24 how the hell is possible disable fb console? nomodeset, nofb, vga=normal, vga=ask, no one of this kernel options works 2013-06-29 20:04:05 ACTION is having a problem with wiki.a.o http://wiki.alpinelinux.org/wiki/File:Acf-mvc.png 2013-06-29 20:04:21 seems this image was deleted some time ago (don't know why) 2013-06-29 20:04:39 when I tried to restore it, it seemed to work, but the image is still missing 2013-06-29 20:04:42 any ideas? 2013-06-29 20:12:04 similar for another ACF image: 05:02, 8 April 2010 K0gen (Talk | contribs | block) deleted page File:Acf-inheritance.png (view/restore) 2013-06-29 20:13:28 looks like K0gen deleted a bunch of files that day :( 2013-06-29 20:13:47 ACTION obviously hasn't looked at this page in a while :) 2013-06-30 14:45:13 ncopa: Could you take a look at my repo? I updated qemu, imagemagick, freetype, logrotate, varnish, json-c and rsyslog – https://github.com/prymeroot/aports