2014-06-01 02:46:30 http://wiki.znc.in/Running_ZNC_as_a_system_daemon 2014-06-01 02:46:32 hm 2014-06-01 02:46:46 I don't think the znc package has an init script yet :o 2014-06-01 07:25:20 Mp5shooter: I don't mind, will do after breakfast 2014-06-01 10:46:29 znc 1.2 is no longer available 2014-06-01 10:56:25 Mp5shooter: if i upgrade it, can you test if it works? 2014-06-01 10:56:45 im not sure what current policy is regarding the 3.0 release. 2014-06-01 10:56:52 so better make sure its in a working state 2014-06-01 11:12:12 we need error.h 2014-06-01 11:54:25 https://paste.xinu.at/lZJdHR/ 2014-06-01 11:54:39 packages from testing with the date of last APKBUILD modification 2014-06-01 11:55:08 only 279 are from this year 2014-06-01 11:55:39 I opt for getting rid of the rest 2014-06-01 11:55:54 clandmeter, ncopa, uggedal ↑ 2014-06-01 11:58:05 barthalion: yeah this crep gives me a headache 2014-06-01 11:58:55 we should basically move everything older than 2 or 3 months 2014-06-01 11:59:09 if someone didn't care to move it yet, he won't anyway 2014-06-01 12:00:04 this way we could limit testing to 152 packages 2014-06-01 12:00:14 and from there we could review the rest 2014-06-01 12:00:15 or have a abuild post hook on failure to auto: git mv $pkgdir ../unmaintained && git commit -c "testing/$pkgname: move to unmaintained" 2014-06-01 12:01:21 it won't hard to write a cron job to do it 2014-06-01 12:03:46 well thats something we can worry about after we cleanup this mess 2014-06-01 12:06:07 barthalion: from that list, can we remove the pkgs thhat have already been build? 2014-06-01 12:07:53 not really 2014-06-01 12:08:35 well, we can 2014-06-01 12:08:57 at least we can skip recently fixed packages 2014-06-01 15:14:42 sure clandmeter 2014-06-01 17:10:46 maybe if there's an easy way to check if the packages are up to date 2014-06-01 17:10:49 and keep those 2014-06-01 17:10:50 ACTION shrugs 2014-06-01 21:00:23 does anyone mind upgrading rutorrent 2014-06-02 08:03:48 good morning 2014-06-02 08:04:06 ncopa: whats your idea about testing repo? 2014-06-02 08:04:29 clandmeter: that we disable the stuff that does not work 2014-06-02 08:04:36 or fix the easy-to-fix things 2014-06-02 08:04:53 there are so much of them 2014-06-02 08:24:00 sisyphean task 2014-06-02 08:24:20 we can get rid of both build failures and unused packages 2014-06-02 11:22:26 ncopa, after fixing the build, and a kernel upgrade... we should be good for new 3.0_rc? 2014-06-02 11:24:12 yup 2014-06-02 11:24:44 still no grsecurity patch for 3.14.5 though 2014-06-02 12:45:41 fabled: i now have 11 patches for nfs-utils totally: http://sprunge.us/KVDP 2014-06-02 12:46:10 the last one fixes the memleak and a segfault on #musl 2014-06-02 12:50:32 ncopa, re attribute_mallo, the #define is used in glibc so that it works on old gcc's 2014-06-02 12:50:52 now you traded portability between libc's to non-portability over gcc/other c-compiler 2014-06-02 12:50:59 yes, for us it's good. not sure if it's upstreamable 2014-06-02 12:51:55 did you verify that posix semantics for basename() work? 2014-06-02 13:02:33 ncopa: sounds good! :-) 2014-06-02 13:04:16 fabled: i think it work good enough. i think its just for error messages/info and such 2014-06-02 13:04:22 and is better than segfaulting... 2014-06-02 13:04:23 hi@all 2014-06-02 13:04:26 hi 2014-06-02 13:17:28 whee! 2014-06-02 13:17:32 i think we have progress! 2014-06-02 13:17:42 mountd: refused mount request from 10.65.66.59 for /var/cache/distfiles (/var/cache/distfiles): unmatched host 2014-06-02 13:19:24 mountd: authenticated mount request from 10.65.66.59:987 for /var/cache/distfiles (/var/cache/distfiles) 2014-06-02 13:19:24 mountd: nfsd_fh: inbuf '10.65.65.0/21 4 \x4a1486007fe7cbe9' 2014-06-02 13:19:24 mountd: nfsd_fh: found 0x720e6f55e0f0 path /var/cache/distfiles 2014-06-02 13:19:38 and there it hanged again 2014-06-02 13:26:11 haha 2014-06-02 13:26:22 but yeah, almost there :) 2014-06-02 13:32:24 there have been musl changes too 2014-06-02 13:32:36 but i am sceptic to nfs-utils 2014-06-02 13:32:50 seems like they do nasty things depending on the libc internals 2014-06-02 13:33:00 which works on glibc... 2014-06-02 14:29:00 can sshfs be used as somewhat/temp alternative to nfs ? 2014-06-02 14:29:25 not really 2014-06-02 17:06:26 where can I get the edge release candidates to play around with? 2014-06-02 17:16:05 i. e. http://dl-5.alpinelinux.org/alpine/v3.0/releases/ 2014-06-02 17:23:23 thanks StarWarsFan 2014-06-02 17:23:34 you're welcome :-) 2014-06-02 17:23:55 the arm dir is empty :( 2014-06-02 17:24:12 and I'm looking at a bored raspberry pi :) 2014-06-02 17:39:32 ScrumpyJack (12:50:58) (@fabled) clandmeter, try http://dev.alpinelinux.org/~tteras/alpine-edge-140529-armhf.rpi.tar.gz 2014-06-02 17:44:55 cool, thanks 2014-06-02 18:18:54 that's strange, I can't seem to boot the ISOs in virtualbox 2014-06-02 19:11:20 ScrumpyJack that is not the iso, just unpack it to fat partition 2014-06-02 19:12:49 ScrumpyJack copy it to a primary fat partition on the sd, (i did not used it myself) 2014-06-02 19:17:19 crow: i'm using virtualbox, it's an old version. i'll upgrade, and it should work 2014-06-02 19:27:21 ScrumpyJack edge is on musl, i am not sure what you have there for rpi. if it was not based on uclibc 2014-06-02 19:36:20 ACTION seems to recall that alpine 2.7 and newer required me to update VirtualBox too 2014-06-02 19:36:24 wouldn't work on the older version 2014-06-02 19:36:33 ScrumpyJack: ^^ 2014-06-02 19:37:15 and the ARM image mentioned above for rpi is based on musl 2014-06-02 19:37:19 just based on the date 2014-06-02 20:19:13 tdtrask: thanks 2014-06-02 20:35:41 virtualbox update fixed that 2014-06-02 20:37:55 np 2014-06-02 20:39:33 when booting the cd, Migrating /var/run to /run failed 2014-06-02 23:34:51 uggedal, unmaintained/libqb is required by testing/corosync 2014-06-02 23:35:01 I hope this is ok ? 2014-06-03 02:07:08 thank you for upgrading rutorrent ncopa :) 2014-06-03 04:30:44 hum 2014-06-03 04:30:57 why is nothing building on edge o.O 2014-06-03 06:16:16 morning 2014-06-03 06:16:28 i reconfiugred the edge builders to halt on first error 2014-06-03 06:16:39 and they are halting in testing repo 2014-06-03 06:17:16 morning 2014-06-03 06:17:20 oh 2014-06-03 06:18:37 nothing in testing or main should depend on anythign in unmaintained 2014-06-03 06:18:44 and nothing in main should depend on anything in testing 2014-06-03 06:20:14 i think something is wrong with lua-aports too 2014-06-03 06:20:51 it should ignore packages with arch="x86 x86_64" 2014-06-03 06:44:49 search.cpan.org is down 2014-06-03 07:46:40 rnalrd: new linux-grsec kernel for edge/v3.0: 3.14.5 2014-06-03 12:01:46 i wonder if i should ignore build errors on edge builders for now... 2014-06-03 12:24:11 fabled: do we have __WORDSIZE, or a replacement? 2014-06-03 12:24:32 in C-headers, getconf or where? 2014-06-03 12:24:58 headers 2014-06-03 12:26:29 seems that bits/user.h and bits/reg.h has it. but that seems non-standard define. 2014-06-03 12:26:33 what's the usecase? 2014-06-03 12:26:40 why not just use sizeof(int)*8 or similar 2014-06-03 12:27:06 err 2014-06-03 12:27:08 sizeof(long) 2014-06-03 12:27:11 or LONGBITS 2014-06-03 12:28:05 pulseaudio complains about it 2014-06-03 12:28:22 ./pulsecore/sample-util.h:60:5: warning: "__WORDSIZE" is not defined [-Wundef] #if __WORDSIZE == 64 || ((ULONG_MAX) > (UINT_MAX)) 2014-06-03 12:29:07 huh. that's kinda stupid test. ULONG_MAX > UINT_MAX should suffice alone 2014-06-03 12:29:47 musl seems to internally use: LONG_MAX > INT_MAX 2014-06-03 13:08:39 drbd had similar issue iirc 2014-06-03 13:30:18 i made edge builders continue on build errors 2014-06-03 19:21:26 is setting up a builder document anywhere? 2014-06-03 19:21:32 I mean lxc musl setup 2014-06-04 07:00:21 openview 2014-06-04 07:00:37 sorry wrong window, btw. morning 2014-06-04 07:01:11 barthalion i am also interesting in lxc builder setup (own repository) 2014-06-04 07:05:17 ncopa: ↑ 2014-06-04 07:12:30 morning 2014-06-04 07:12:52 i have some notes from when i set up the official build servers (build-3-0-x86/x86_64) 2014-06-04 07:12:59 http://wiki.alpinelinux.org/wiki/User:Ncopa/buildserver 2014-06-04 07:18:52 ncopa thx 2014-06-04 07:19:39 crow: do you want upload the packages some place too? 2014-06-04 07:19:49 or is just autobuilding things enough? 2014-06-04 07:20:02 i'm not sure you can disable the upload part 2014-06-04 07:20:09 but i could make it configurable if needed 2014-06-04 07:43:52 ncopa: I want to have failing testing here 2014-06-04 07:44:15 we hsould have that on edge builders yes 2014-06-04 07:44:29 but i disabled it for now, til i have done the v3.0 release 2014-06-04 07:45:04 mkay then 2014-06-04 07:51:44 ncopa curretly only for testing, without upload, so it would be nice to have it in future configurable 2014-06-04 08:44:16 ncopa: any idea why lxc doesnt include lxc-device? 2014-06-04 08:49:48 dunno 2014-06-04 08:51:40 ncopa: its because of python3 bindings 2014-06-04 08:51:51 can i add those? 2014-06-04 09:00:02 clandmeter: if you split it so we dont get a hard dep of python3 yes 2014-06-04 09:00:26 i wonder if it will replace lxc-ls and others with python variants if you enable python 2014-06-04 09:03:24 ncopa: you are right 2014-06-04 09:03:31 it will replace lxc-ls with python version 2014-06-04 09:06:14 https://github.com/lxc/lxc/blob/master/src/lxc/lxc-device 2014-06-04 09:06:32 it only uses the liblxc i think 2014-06-04 09:12:03 should be relatively simple to write that tool in either C or lua 2014-06-04 10:44:33 fabled: anyone ever got to add those debug parts to musl? 2014-06-04 10:44:58 clandmeter, i had few bits for x86 only, but it's not really done. 2014-06-04 10:45:20 i've made some noise about it on the musl mailing list 2014-06-04 10:55:44 debug stuff for oscam in my case are needed so i hope that part will work :) 2014-06-04 11:35:08 im actually also worried about non-working valrind 2014-06-04 11:35:21 valgrind is kinda essential to find memleaks 2014-06-04 11:35:50 kaniini: xen does not work with musl at all right? 2014-06-04 11:36:13 maybe we should simply drop building alpine-xen.iso til its fixed? 2014-06-04 11:44:57 can anyone confirm that alpine-xen 3.0 is broke? 2014-06-04 11:55:02 i'm tagging 3.0.0_rc5 now 2014-06-04 12:10:23 if alpine-vanilla and alpine 3.0.0_rc5 isos boot I'll just tag v3.0.0 2014-06-04 12:11:32 ncopa: I can give it a try later 2014-06-04 12:13:18 ncopa: were can I find the alpine-xen isos for 3.0? 2014-06-04 12:14:38 http://nl.alpinelinux.org/alpine/v3.0/releases/x86_64/ 2014-06-04 12:14:43 i didnt build rc5 2014-06-04 12:14:48 i think it fails 2014-06-04 12:15:00 kaniini is working on it i think 2014-06-04 12:22:28 OK, sorry for that, I've been away half of the past month :( 2014-06-04 12:23:49 as i understood kaniini, the problem appears to be when building 32bit binaries with gcc -m32 on a 64bit host 2014-06-04 12:24:02 for some reason it ends up pulling in the 64bit headers 2014-06-04 12:24:17 so it might be a gcc/toolchain problem 2014-06-04 12:24:55 ok, i think I just tag 3.0.0 2014-06-04 12:25:36 mmm, maybe we are not correctly specifing the 32bit emulation in the linker 2014-06-04 12:25:49 kaniini: drop me a line if you want to comment on the Xen issues 2014-06-04 12:26:56 ok, i need to write releasenotes 2014-06-04 12:34:14 wooo 2014-06-04 13:05:12 ok 2014-06-04 13:05:18 what more can we put in release notes? 2014-06-04 13:13:21 ncopa, compared to what? you have draft somewhere? 2014-06-04 13:14:10 https://dpaste.de/ZqR3 2014-06-04 13:14:26 if there is anything you gusy think should be mentioned... 2014-06-04 13:14:29 musl ofcourse 2014-06-04 13:14:32 and abi incompat 2014-06-04 13:14:44 are there other major upgrades among the packages? 2014-06-04 13:14:51 or any specially interesting new packages? 2014-06-04 13:16:12 asterisk is now 12.x 2014-06-04 13:16:34 it's stil 11.x in 2.7-stable 2014-06-04 13:16:54 ah it's mentioned 2014-06-04 13:17:11 ncopa, just pushed roundcube 1.0 this morning 2014-06-04 13:19:02 new nginx? 2014-06-04 13:19:36 ffmpeg + mpv? 2014-06-04 13:19:44 vim 7.4? 2014-06-04 13:20:22 ncopa: I guess you can merge my commits into one entry 2014-06-04 13:20:28 does freeswitch works ? 2014-06-04 13:20:36 Idk why I made 70 commits with my personal address 2014-06-04 13:21:50 fcolista, freeswitch does not build, and is not available for 3.0 2014-06-04 13:21:53 it's in release notes 2014-06-04 13:22:15 ncopa, new firefox? 2014-06-04 13:22:22 mentioned too 2014-06-04 13:22:24 i'm slow today 2014-06-04 13:22:34 ok 2014-06-04 13:23:28 freeswitch is in good company like erlang 2014-06-04 13:25:21 should we mention about difficulties with gdb? 2014-06-04 13:25:39 might be an idea to mention 2014-06-04 13:25:45 ncopa, one thing to add is Full RELRO build 2014-06-04 13:26:17 new varnish? 2014-06-04 13:27:25 yes i was thinking of gdb 2014-06-04 13:27:33 limited gdb support or soemthing 2014-06-04 13:27:39 varnish indeed 2014-06-04 13:27:50 didnt we have full RELRO build before? 2014-06-04 13:28:58 what should i say about gdb? 2014-06-04 13:30:05 i think fabled can descibe that best. 2014-06-04 13:30:12 ncopa, i think we were supposed to have RELRO but the gcc patch never worked right earlier 2014-06-04 13:31:05 re: gdb, "musl libc has still limited support for debugging, so gdb backtraces might be incomplete" or similar 2014-06-04 13:32:55 update: https://dpaste.de/QGM9 2014-06-04 13:33:00 i added varnish 2014-06-04 13:33:08 full relro support 2014-06-04 13:33:21 mention incomplete debugging 2014-06-04 13:33:29 i think thats okish 2014-06-04 13:34:08 wow, fabian_a has more commits than both fabled and barthalion 2014-06-04 13:35:05 not really 2014-06-04 13:35:14 oh 2014-06-04 13:35:15 :) 2014-06-04 13:35:17 yup 2014-06-04 13:35:18 I'm twice on the list 2014-06-04 13:35:22 you changed email 2014-06-04 13:35:55 ncopa, should we mention "shorter support" 2014-06-04 13:36:02 right, this 2014-06-04 13:36:08 oh thats true 2014-06-04 13:36:40 otherwise i suppose it's good to go 2014-06-04 13:36:49 oh 2014-06-04 13:36:50 arm images 2014-06-04 13:37:02 "no arm yet?" 2014-06-04 13:37:38 or "arm port has been created, but is still experimental and not included in release builds" ? 2014-06-04 13:38:08 yeah 2014-06-04 13:38:12 that sounds more positive 2014-06-04 13:38:53 A port for ARM has been created, but is still experimental and not included in release builds. 2014-06-04 13:39:34 update: https://gist.github.com/ncopa/e9697f55ccc22ea10763 2014-06-04 13:42:15 i wrapped long lines 2014-06-04 13:42:44 this will go to distrowatch, so anything we could say about xorg/desktop use? 2014-06-04 13:43:01 too bad we dont have new logo yet... 2014-06-04 13:45:04 yeah 2014-06-04 13:45:08 i'm good with that 2014-06-04 13:47:59 looks like ncopa still did about 50% of all changes 2014-06-04 13:48:02 good job! 2014-06-04 13:49:13 yeah, ncopa is unbeatable :) 2014-06-04 13:49:54 i cheat ;) 2014-06-04 13:50:14 i think most of the commits are done by a script... 2014-06-04 13:50:58 if anything gets complicated i push it to fabled ;) 2014-06-04 13:52:17 oh we should maybe have mentioned that we dropped vserver support 2014-06-04 13:53:26 ncopa maybe to mention no properitie driver packages 2014-06-04 13:53:40 thats not really any news 2014-06-04 13:53:42 lxc is mentioned 2014-06-04 13:53:59 but yeah. vserver has been removed for good is a major change 2014-06-04 13:54:25 http://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases needs update too 2014-06-04 13:54:35 and front page + about re: musl 2014-06-04 13:54:52 and the download link 2014-06-04 13:54:58 ncopa maybe not, but if on distrowatch and mentioned xorg/desktop use ppl will ask for properierts drivers (i guess) 2014-06-04 13:55:09 right.. 2014-06-04 13:55:16 we should mention that in about page actually 2014-06-04 13:55:23 updating release page 2014-06-04 13:55:32 yes 2014-06-04 13:57:29 ncopa maybe that unmaintened packages and not builded packages are moved to unmaintened repo and are waitinig for "you" guys :D 2014-06-04 13:58:15 we should send an 'Alpine 3.0' -cake to ncopa ;) 2014-06-04 14:00:12 we don't expect 3.0 to have a long life, don't we? 2014-06-04 14:00:24 supported for 1y? 2014-06-04 14:00:37 1y max 2014-06-04 14:00:56 i think i mentioned that in release notes 2014-06-04 14:01:15 yes, 1y support was mentioned 2014-06-04 14:01:26 though 2014-06-04 14:01:32 i'd go wiht 11 months 2014-06-04 14:01:36 until next spring release is out 2014-06-04 14:01:40 hopefully in May 1st 2014-06-04 14:01:52 ~1 year 2014-06-04 14:01:57 i've set 2015-05-01 2014-06-04 14:01:58 yes 2014-06-04 14:02:29 done 2014-06-04 14:02:40 rnalrd, branch date is off by month 2014-06-04 14:02:47 it's june already... 2014-06-04 14:02:53 lol 2014-06-04 14:02:53 yes, we are late. 2014-06-04 14:02:54 yes 2014-06-04 14:05:37 should be good 2014-06-04 14:06:01 uff 2014-06-04 14:06:03 top link 2014-06-04 14:13:01 front page updated 2014-06-04 14:20:26 about page updated (but needs more fixes i think) 2014-06-04 14:20:50 i have been reading about static web page generators 2014-06-04 14:20:55 like jekyll, nikola etc 2014-06-04 14:21:19 nikola \o/ 2014-06-04 14:21:20 there are also some nice things that suckless.org uses, which i like alot 2014-06-04 14:21:29 in general 2014-06-04 14:21:41 i love the concept of generated static pages 2014-06-04 14:21:46 generated with gnu make 2014-06-04 14:22:25 there was some other that i think looked slightly better than nikola, i think it was the top rated python 2014-06-04 14:22:36 http://staticsitegenerators.net/ 2014-06-04 14:22:55 http://blog.getpelican.com/ 2014-06-04 14:22:58 but 2014-06-04 14:23:13 i think its should be relatively easy to roll our own 2014-06-04 14:23:42 looking at the suckless.org generator 2014-06-04 14:24:01 the site has relatively static content 2014-06-04 14:24:13 i mean our alpinelinux.org site has mostly static content 2014-06-04 14:24:28 when there are needed for change, we could simply rebuild the pages with make 2014-06-04 14:25:12 i think making news articles is simple 2014-06-04 14:25:19 and index those 2014-06-04 14:25:45 download links etc should be faily straightforward too 2014-06-04 14:25:58 i think package browser could be static pages too, generated with make 2014-06-04 14:26:18 there would be a bit work to import the current stuff 2014-06-04 14:26:32 problem might be forums 2014-06-04 14:26:45 well, a 'challenge' 2014-06-04 14:40:51 I see two major reasons to pick nikola 2014-06-04 14:40:59 it can do more than pelican 2014-06-04 14:41:21 and I know its current maintainer 2014-06-04 14:41:36 and about forums… 2014-06-04 14:41:46 maybe we should host Q&A, not traditional bbs 2014-06-04 14:43:45 i was thinking for comments 2014-06-04 14:43:54 we could use discuss like nikola 2014-06-04 14:44:07 or we could use bbs for comments like phoronix.com doe it 2014-06-04 14:44:17 fluxbb seems okish 2014-06-04 14:47:08 so when will edge get its updated packages? 2014-06-04 14:50:13 they should be there now 2014-06-04 14:50:28 hm 2014-06-04 14:50:35 maybe the mirror I'm using hasn't updated 2014-06-04 14:50:43 i switched back to 'skip errors' mode for now 2014-06-04 14:51:17 I switched from dl-4 to nl.a.o 2014-06-04 14:51:20 and I get updates now 2014-06-04 14:51:33 looks like dl-4 is still on v3.0.0_rc3-10-g801eaa8 2014-06-04 14:51:38 hm 2014-06-04 14:51:50 ok i'll check that tm 2014-06-04 14:53:06 ok 2014-06-04 14:54:56 fluxbb will do as well, but it doesn't differ much from current forums 2014-06-04 14:55:01 I see no advantage here 2014-06-04 14:56:42 it would be change for sake of changing 2014-06-04 15:03:14 well, i dunno if it would be worth keeping drupal only for the forums 2014-06-04 15:03:27 it would be the only thing drupal would do 2014-06-04 15:24:04 I still opt for something like stackoverflow 2014-06-04 19:27:18 I think there's something wrong with roundcubemail on musl 2014-06-04 19:27:31 it says it can't read the config files and it can't write to tmp 2014-06-04 19:27:35 which did not happen before 2014-06-04 19:40:21 ncopa: static site gen with shell and make is quite easy: https://github.com/uggedal/uggedal.com 2014-06-04 19:41:23 cheated with python for markdown rendering with integrated syntax highlighting though.. 2014-06-04 20:24:07 so no alpine-xen 3.0 ? perhaps we could have a 64bit version and postpone the 32bit version till later? 2014-06-04 20:29:55 (if building a 32bit version is the problem) 2014-06-04 20:46:02 perhaps the setup-xen-dom0 should be removed from the other builds 2014-06-05 00:27:28 used the manual sql database setup script that comes with roundcube to get the database set up and all was well after that 2014-06-05 00:27:36 idk why roundcubemail installer was all weird o.O 2014-06-05 06:05:33 morning 2014-06-05 06:31:17 uggedal: yes, rolling something yourself with markdown is fairly simple 2014-06-05 06:31:26 "Generated with POSIX shell." 2014-06-05 06:31:28 :-D 2014-06-05 06:31:31 i love that one :) 2014-06-05 07:10:46 morning 2014-06-05 07:14:37 so what's the issue with alpine-xen? 2014-06-05 07:22:09 kaniini knows details 2014-06-05 07:22:22 but as i understand, there is a problem how it got built 2014-06-05 07:22:31 non-trivial to find out apparently 2014-06-05 07:49:54 so 2014-06-05 08:05:58 basically, the issue is 2014-06-05 08:06:15 hvmloader is built wrongly, it should use gcc headers but it uses musl's headers 2014-06-05 08:07:11 basically, -nostdinc -isystem $gcc_include_dir is what we need to add to cflags for hvmloader 2014-06-05 08:07:16 otherwise the xenpv stuff works fine 2014-06-05 08:07:51 hvm is usable on alpine 3, but at present, requires an hvmloader built on alpine 2.7 2014-06-05 08:58:37 ncopa: not really a bug 2014-06-05 08:58:40 kaniini: do you think you could file a bug on bugs.a.o with the details for xen? 2014-06-05 08:58:49 ncopa: kernel mitigates such issues on its own 2014-06-05 08:59:17 ncopa: https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/tree/kernel/sys.c?id=refs/tags/v3.14.5#n415 2014-06-05 08:59:37 barthalion: i kinda thought that its not the job of bash to maintain security for that... 2014-06-05 09:00:49 I applied this patch in Arch only to keep my inbox "zomg bash is insecure" free 2014-06-05 09:01:02 yeah.. 2014-06-05 09:01:15 lots of those "sec" issues are in that category... 2014-06-05 15:24:12 is there a virtualbox package for alpine linux? 2014-06-05 15:28:04 i'm quite happy making one. I have no experience of building against musl. Does is work as drop in for glibc? Will I need to build new makefiles? I thought it would be nice to run virtualbox in a alpine dom0 (although I'm not sure it makes much sense) 2014-06-05 15:29:49 perhaps i should start with a small, more trivial package. I use sc, I could try packaging that. 2014-06-05 15:30:00 time to look at the apk docs 2014-06-05 15:34:54 virtualbox can be quite difficult 2014-06-05 15:35:05 but possible 2014-06-05 15:37:57 ScrumpyJack: running virtualbox in a Xen Dom0 is completely nonsense 2014-06-05 16:25:38 royger: I have a feeling you might be right, but can you elaborate? 2014-06-05 16:29:08 ScrumpyJack: when running as Dom0, you don't have access to any CPU virtualization extensions from the CPU (because they are used by Xen), so you would fallback to software emulation which is extremly slow 2014-06-05 16:29:28 ScrumpyJack: also, why would you do it when you can just launch a regular Xen guest? 2014-06-05 16:37:30 because my (very old) hvm guest OS doesn't pick up my change of CD during installation (when i use the phyisical drive or when I block-attach isos). I have a prebuild vbox of the old OS, so i thought, why not load that. 2014-06-05 16:39:47 ScrumpyJack: I think this has been fixed in recent versions of Xen, 4.3 should detect those media changes just fine 2014-06-05 16:40:18 ScrumpyJack: if you have to use vbox, just boot a regular Linux without Xen, or else it's going to be extremly slow 2014-06-05 16:47:09 I had virtualbox running in a linux pv guest so I could pull a vagrant docker build to play with LXC inside that. all seemed fine. 2014-06-05 16:49:30 that's several LXCs on linux running in vbox on a linux pv on xen with alpine dom0 2014-06-05 16:51:23 ScrumpyJack: that seems awfully complicated to me, but anyway, why didn't you just installed docker inside of another linux pv guest? 2014-06-05 16:55:25 Just playing around, checking out vagrant. I like the start, play, teardown side of it. 2014-06-05 16:58:01 royger, please note that I'm using alpine-xen, and that I love it :) 2014-06-05 20:11:43 ncopa: yes, but i will probably just fix it in a day or two 2014-06-05 21:45:11 openldap is segfaulting on musl 2014-06-05 23:31:07 this is freeradius 2014-06-05 23:31:13 after upgrading openssl 2014-06-05 23:31:14 libssl version mismatch. Built with: 1000107f Linked: 1000108f 2014-06-05 23:31:19 won't start now 2014-06-06 04:58:03 Mp5shooter, oh yes, that needs to be patched. the freeradius ssl version check is broken 2014-06-06 04:58:17 oh 2014-06-06 04:58:36 do you know if openldap is broken too? it's segfaulting for me 2014-06-06 07:03:20 morning 2014-06-06 14:19:42 have a good weekend all 2014-06-06 14:25:40 ScrumpyJack: you too 2014-06-06 23:48:11 does anyone mind rebuilding freeradius package against the latest libssl in repos so it can start 2014-06-07 06:47:25 we should probably add man compression support to abuild 2014-06-07 06:47:46 scandium:~# find /usr/share/man -type f | wc -l 2014-06-07 06:47:46 4023 2014-06-07 06:47:46 scandium:~# find /usr/share/man -type f -name \*.gz | wc -l 2014-06-07 06:47:47 25 2014-06-07 13:03:05 ugh, why webkitgtk installation is so slow? 2014-06-07 13:04:02 and now it blasted 2014-06-07 13:50:04 ncopa: patch for abuild man page compression: http://sprunge.us/PYKb 2014-06-07 15:17:19 does anyone mind rebuilding freeradius package against newest libssl in edge? it won't start 2014-06-07 20:30:04 dl-4 is missing the 3.0 x64 iso :-/ 2014-06-07 20:31:49 yeah there's something wrong with dl-4 2014-06-07 20:31:56 I think ncopa was looking into it 2014-06-07 20:32:03 ah ok 2014-06-07 20:32:05 I switched to nl.a.o 2014-06-07 20:32:06 nvm then 2014-06-07 20:32:14 because packages aren't getting updated eitehr 2014-06-07 20:32:15 ACTION runs dl-5 2014-06-07 20:32:23 oh nice 2014-06-07 20:33:11 just got the error 'cause i wanted to download the 3.0 x64 iso from alpinelinux.org 2014-06-07 20:33:44 what's the 3.0 default kernel version? 2014-06-07 20:33:54 ACTION shrugs 2014-06-07 20:33:59 I don't pay attention to that stuff :p 2014-06-07 20:34:35 http://i.imgur.com/1RviT44.png 2014-06-07 20:34:42 that's why i'm asking 2014-06-07 20:35:19 oh man 2014-06-07 20:35:21 that happened to me 2014-06-07 20:35:24 when I upgraded to musl 2014-06-07 20:35:33 I couldn't fix it and ended up scrapping the box 2014-06-07 20:35:49 i'm sitting in front of an hp dl320e g8 2014-06-07 20:36:02 not booting from harddisk, not booting ubuntu ... 2014-06-07 20:36:06 kinda worthless piece of crap 2014-06-07 20:36:07 :S 2014-06-07 20:36:20 oh 2014-06-07 20:36:23 I didn't see that firmware bug line 2014-06-07 20:36:32 that didn't happen to me but everything under it did 2014-06-07 20:37:06 yeah, stuff like this sucks 2014-06-07 20:38:00 if my girlfriend could suck like this stuff 2014-06-07 20:38:02 hell of a job 2014-06-07 21:31:42 :P 2014-06-07 21:31:58 and yeah 2014-06-07 21:32:01 it was pretty frustrating 2014-06-07 21:32:04 ruined many days for me 2014-06-07 21:38:32 tell me about it 2014-06-07 21:57:54 grimeton: you need modify /etc/mkinitfs/features.d/ files, just move the apk-new files into place 2014-06-07 21:58:00 +to 2014-06-07 21:58:17 kaniini: eeer what? 2014-06-07 21:58:19 oh wait 2014-06-07 21:58:24 is this from ISO boot? 2014-06-07 21:58:28 kaniini: yes 2014-06-07 21:58:32 in that case, you need 3.x iso 2014-06-07 21:58:38 i guess 2014-06-07 21:58:59 this is the alpine-mini-3.0.0-x86_64.iso 2014-06-07 21:59:14 weird 2014-06-07 21:59:20 na not weird 2014-06-07 21:59:21 HP 2014-06-07 22:00:02 it's ok 2014-06-07 22:00:12 i recently wrote a keygen for HP lights out 100i 2014-06-07 22:00:22 so that you can use the kvm without paying for it ;p 2014-06-07 22:00:24 you don't happen to have one for the ilo 4? 2014-06-07 22:00:34 unfortunately i do not have an iLO 4 machine 2014-06-07 22:00:40 or time to look at it in idapro 2014-06-07 22:00:41 ;) 2014-06-07 22:00:46 :) 2014-06-07 22:00:51 nvm. seen a few around 2014-06-07 22:01:34 that stupidity is why supermicro is the most popular server vendor even though everything they make is awful 2014-06-07 22:01:45 because other servers are more painful 2014-06-07 22:02:04 it takes longer to wait while the hp server goes through the bios than it takes to boot the main server later (if it boots) 2014-06-07 22:02:22 "oh, you wanted to use the KVM on your iLO? that will be $500 please" 2014-06-07 22:02:28 yeah 150 2014-06-07 22:02:31 or, with dell, it is like 2014-06-07 22:02:32 129 actually 2014-06-07 22:02:41 "buy an idrac enterprise license" 2014-06-07 22:02:49 and it adds extra $ to the thing 2014-06-07 22:03:04 i usually don't buy hp or dell or ibm servers anymore 2014-06-07 22:03:07 for lack of a better word, i would call these things antifeatures 2014-06-07 22:03:32 someone obviously went in and wrote code to make the KVM intentionally not work unless a key was installed 2014-06-07 22:03:45 and wrote unit tests or whatever to prove it worked (ha ha, yeah right) 2014-06-07 22:03:54 kaniini: yeah ... it's basically a linux system running on a ppc board 2014-06-07 22:04:07 afaik 2014-06-07 22:04:18 lo100 is arm-based 2014-06-07 22:04:26 yeah the ibm ones are ppc based 2014-06-07 22:04:36 you can clearly see the big ppc cpu :D 2014-06-07 22:04:39 powerpc something 2014-06-07 22:04:40 same chip as the supermicros even 2014-06-07 22:04:45 doesn't even carry a heatpipe or something 2014-06-07 22:04:56 ibm calls them RMMI iirc 2014-06-07 22:05:17 oh, one of those big ceramic packages? 2014-06-07 22:05:21 yeah 2014-06-07 22:05:32 how retro ;) 2014-06-07 22:05:34 sometimes it even has the golden platter 2014-06-07 22:07:23 http://en.wikipedia.org/wiki/IBM_Remote_Supervisor_Adapter#mediaviewer/File:Rsaii.jpeg 2014-06-07 22:07:52 the small board at the top 2014-06-07 22:07:59 haha naturally 2014-06-07 22:09:19 brb 2014-06-09 13:36:43 i really miss the support for multiple git branches feature in redmine 2014-06-09 14:26:04 ohi 2014-06-09 14:28:57 hi 2014-06-09 14:29:35 WwQCn4HqahlD: what is the clone url? 2014-06-09 14:30:18 ncopa: you mean this? https://github.com/stef/aports.git 2014-06-09 14:30:54 yup 2014-06-09 14:30:56 thanks 2014-06-09 14:31:41 i have also packaged other stuff, but musl seems to break some things 2014-06-09 14:31:48 did you try build the things? 2014-06-09 14:31:50 https://github.com/stef/aports/commit/7676832921a0dc4e4b40be41d45c9e7bd62dd507 2014-06-09 14:31:54 i packaged radare2 and radare2capstone 2014-06-09 14:31:58 that looks broke 2014-06-09 14:32:14 why? 2014-06-09 14:32:19 ah 2014-06-09 14:32:19 no 2014-06-09 14:32:34 i didnt see the patch file rename 2014-06-09 14:34:42 i also notified upstream already about my musl patches to zfs 2014-06-09 14:34:50 great! 2014-06-09 14:34:52 thanks1 2014-06-09 14:36:14 hey ncopa is it still possible to have strongswan backported to 2.6.6? 2014-06-09 14:40:19 Mp5shooter: everything is possible. the impossible just takes longer time ;) 2014-06-09 14:40:29 heh :p 2014-06-09 14:40:39 Mp5shooter: i would appreciate if you filed a bug for it: http://bugs.alpinelinux.org 2014-06-09 14:40:46 sure 2014-06-09 14:40:48 with 2.6.* as target 2014-06-09 14:40:57 so i dont forget 2014-06-09 14:41:40 ok 2014-06-09 14:43:53 filed :) 2014-06-09 14:44:02 thank you very much 2014-06-09 14:58:40 thanks ncopa! 2014-06-09 14:58:49 much appreciated :) 2014-06-09 14:58:53 np 2014-06-09 14:59:06 off to work now, see you later 2014-06-09 18:05:06 did i hear zfs? 2014-06-10 06:47:31 morning 2014-06-10 08:52:06 so, a common question people keep asking me is 2014-06-10 08:52:31 are we going to switch openssl out for some dodgy libressl port done on github (the one of choice changes every day) 2014-06-10 08:52:47 simple answer is: no, we will wait until libressl-portable is actually released by openbsd 2014-06-10 08:54:49 the other, dumber question is, "since you guys modified openssl to remove the debug malloc and memory caching code in libssl, will you also be forking openssl or maintaining a large patchset on it" -- answer is: no, we will just keep it as it is for now (unless some huge other screwup is found, at which point we will mitigate it with patching until libressl is available) 2014-06-10 08:57:49 also, will alpine do it's own libressl port -- answer to that is: not unless there is significant delays in a libressl-portable release, in which case we might wind up doing such a thing, but it seems unlikely 2014-06-10 09:38:21 files from v2.5.4-302-ge5b2b69 uploaded 2014-06-10 13:03:05 hmpf. firefox-30.0 is buggy 2014-06-10 13:47:31 LXC question. I have no NIC when creating a alpine LXC from within alpine with the default templates. before i trawl through the LXC documentation, is there anything obvious that comes to mind? 2014-06-10 13:48:44 you need to configure some kind of nic 2014-06-10 13:48:57 you can for example add this to /etc/lxc/default.conf: 2014-06-10 13:49:04 lxc.network.type = veth 2014-06-10 13:49:04 lxc.network.flags = up 2014-06-10 13:49:04 lxc.network.link = br0 2014-06-10 13:49:06 ScrumpyJack: did you check our wiki? 2014-06-10 13:50:35 i have checked the wiki for a number of things, but not for LXC yet. I'll update the wiki with my steps 2014-06-10 13:50:38 if needed 2014-06-10 13:51:23 everything is there to setup lxc 2014-06-10 13:51:33 including what ncopa said 2014-06-10 13:52:11 its so simple, even i could set it up ;-) 2014-06-10 13:52:46 Nothing links to LXC on the wiki, I'll add a link to the virtualization section 2014-06-10 13:54:18 http://wiki.alpinelinux.org/wiki/LXC 2014-06-10 13:54:50 yup, but nothing links to that page (alt-shift-j) 2014-06-10 13:54:55 try iy 2014-06-10 13:55:02 try it, even 2014-06-10 13:55:55 i dont get it? 2014-06-10 13:58:01 ScrumpyJack: i got it 2014-06-10 13:58:21 clandmeter: on left menu, there is a "what links here" 2014-06-10 13:58:43 I've just added it to the Virtualization section 2014-06-10 13:58:46 and looks like ScrumpyJack fixed it :) 2014-06-10 13:58:50 thanks! 2014-06-10 13:59:36 I really want to update the xen docs too, but I wonder if it might be better to wait for alpine-xen-3 2014-06-10 14:07:13 ScrumpyJack: good catch :) 2014-06-10 14:53:55 please can someone type "useradd" on CLI 2014-06-10 14:53:57 ? 2014-06-10 14:54:08 see what happens here: http://hastebin.com/ohehojutit.vhdl 2014-06-10 14:54:32 it answers as being addgroup 2014-06-10 14:54:36 s/useradd/adduser 2014-06-10 14:55:03 not really, it answers correctly 2014-06-10 14:55:14 busybox adduser doesn't support -m, just as error msg says 2014-06-10 14:55:15 mine is crazy =] 2014-06-10 14:55:46 yeap.. i am aware of it. The problem is not the -m is the fisrts answers 2014-06-10 14:56:06 it says: Usage: addgroup [-g GID] [USER] GROUP 2014-06-10 14:56:32 but i am not calling addgroup :) 2014-06-10 14:57:19 but when i give it a wrong option it says "adduser" correctly 2014-06-10 14:58:10 ah, right 2014-06-10 14:58:28 same happens if you do "busybox adduser" and 2014-06-10 14:58:33 "busybox addgroup"? 2014-06-10 14:58:52 yeap 2014-06-10 14:58:55 :) same 2014-06-10 14:59:22 do you happen do remember what version we ship? I'm not close to my laptop atm 2014-06-10 14:59:23 http://hastebin.com/agugarijow.vhdl 2014-06-10 15:00:05 hm, 1.21.1 2014-06-10 15:00:10 yeap 2014-06-10 15:00:28 it was probably fixed in 1.22 or 1.22.1 as I can't reproduce it 2014-06-10 15:00:33 I'll take a look later today 2014-06-10 15:00:38 ;) 2014-06-10 15:00:51 it won't hurt if you create a ticket on our bugtracker 2014-06-10 15:00:54 AL 2.7.8 2014-06-10 15:00:55 just in case I forget :) 2014-06-10 15:01:03 ok.. i'll do it 2014-06-10 15:02:09 thx 2014-06-10 15:02:54 ncopa: have you seen #3008? 2014-06-10 15:46:36 works for me in 2.7.8, 1.21.1 2014-06-10 15:47:16 http://bugs.alpinelinux.org/issues/3024 2014-06-10 15:47:20 barthalion: ^^ 2014-06-10 15:51:00 LXC section of the wiki doesn't work if your alpine-linux is a domU pv guest. probably due to the bridging of a bridged interface. If i get it working, I'll create a new page for LXC in alpine-domU pv 2014-06-10 15:56:02 alacerda: thanks again 2014-06-10 16:02:57 most of the wiki docs related to virtualisation are alpine specific, perhaps they should start with and relate to alpine-xen? Folks would probably download alpine-xen, and then look for documentation, only to find the alpine specific docs available. does that make sense? 2014-06-10 16:04:21 (for exemple, the docs say "apk add xen" etc etc) 2014-06-10 20:30:19 question regarding edge vs. 3.0 2014-06-10 20:30:38 is it possible to change an edge machine to use v3.0 repo? 2014-06-10 20:31:16 the box was migrated to musl-edge as it was available 2014-06-10 20:31:36 but now as 3.0 is out, i'd like to change it to use the v3.0 pkg-repo 2014-06-11 06:19:28 barthalion: i saw #3008 2014-06-11 06:20:55 StarWarsFan|afk: should not be any problem replace edge with v3.0 2014-06-11 06:21:04 just do that in repositories and: apk upgrade -U -a 2014-06-11 06:32:06 Moinmoin 2014-06-11 06:38:39 ncopa: ok, thx 2014-06-11 06:40:09 morning :) 2014-06-11 07:22:50 Morning 2014-06-11 07:25:18 morning 2014-06-11 07:25:41 ncopa: i see you tried to get freeswitch build on musl? 2014-06-11 07:26:17 yes 2014-06-11 07:26:29 i got it build with default config from git 2014-06-11 07:26:42 but some of the bundled 3rd party libs does not build 2014-06-11 07:26:59 and needs to be patched in a non-standard way... 2014-06-11 07:28:04 yes i hear they are fixing those in trunk 2014-06-11 07:28:24 making it dep on external libs 2014-06-11 07:28:49 got this error when updating to 3.0 from 2.7.8. trivial? 2014-06-11 07:28:49 ERROR: musl-utils-1.1.2-r0: trying to overwrite sbin/ldconfig owned by uclibc-utils-0.9.33.2-r27 2014-06-11 07:29:28 same for ldd, getconf and getent 2014-06-11 07:29:32 you should be able to remove uclibc-utils 2014-06-11 07:29:41 if upgrade is ok. 2014-06-11 07:29:49 apk fix 2014-06-11 07:29:53 should fix it aferwards 2014-06-11 07:31:42 apk Seg faults now 2014-06-11 07:32:17 rebooting 2014-06-11 07:32:24 wait! 2014-06-11 07:32:49 youre in a half performed upgrade state 2014-06-11 07:32:55 half is uclibc and half is musl 2014-06-11 07:33:03 thats why it segfailts 2014-06-11 07:33:10 you should use apk-static 2014-06-11 07:33:31 as described here: http://alpinelinux.org/edge-musl 2014-06-11 07:33:38 does wget work? 2014-06-11 07:35:11 yup. but don't worry. I'll read that page and start again (it's a alpine xen pvvsnapshot) 2014-06-11 07:35:22 phew :) 2014-06-11 07:35:25 basically 2014-06-11 07:35:38 apk add apk-tools-static 2014-06-11 07:35:52 apk.static upgrade -U -a --no-self-upgrade 2014-06-11 07:36:07 might be an idea to apk add busybox-static too before starting 2014-06-11 07:38:25 i'll add apk static for now and try 2014-06-11 07:38:32 round 2 2014-06-11 07:40:53 ERROR: musl-utils-1.1.2-r0: trying to overwrite sbin/ldconfig owned by uclibc-utils-0.9.33.2-r27 2014-06-11 07:40:58 thats a bug actually 2014-06-11 07:41:00 we should fix that 2014-06-11 07:41:02 on reboot, i get Loading modules ...modprobe: can't change directory to '3.10.40-0-grsec': No such file or directory 2014-06-11 07:41:21 did you do apk fix after the upgrade? 2014-06-11 07:41:34 nope 2014-06-11 07:41:42 round 3 ... 2014-06-11 07:41:46 wait 2014-06-11 07:41:52 i think its other problem this time 2014-06-11 07:42:09 i guess you specify a kernel in xen config? 2014-06-11 07:42:23 i think this does not match the kernel installed in guest 2014-06-11 07:42:33 aha, yes, of course. 2014-06-11 07:43:22 anyway, I have round 3 available already. anything you want me to try? 2014-06-11 07:44:04 ACTION gets some coffee 2014-06-11 07:44:30 maybe try espresso? :) 2014-06-11 07:45:01 i would like to get pvgrub tested and documented 2014-06-11 07:45:11 it does not look that hard to get working 2014-06-11 07:48:35 found some tools/utils, wondering if would be useful for AL 2014-06-11 07:48:36 http://bitmath.org/code/ 2014-06-11 07:49:23 gesture control via webcam requested here, http://bugs.alpinelinux.org/issues/1452 2014-06-11 08:00:31 A super lightweight C JSON parser - https://github.com/quartzjer/js0n 2014-06-11 08:19:32 ncopa: do you have a apk package of pvgrub? 2014-06-11 08:23:06 http://sprunge.us/bRZZ 2014-06-11 08:23:17 looks like its shipped with 'xen' package 2014-06-11 08:23:26 maybe we should ship that in separate package? 2014-06-11 08:23:51 xen-pvgrub 2014-06-11 08:23:56 is that what others do? 2014-06-11 08:24:57 well, Debian doesn't ship it at all 2014-06-11 08:26:35 so whatever you do it will be fine 2014-06-11 08:26:45 redhat/suse etc? 2014-06-11 08:27:10 RH dropped xen support 2014-06-11 08:28:04 no idea about suse 2014-06-11 08:28:16 kaniini: do you mind if i move pvgrub files to a xen-pvgrub subpackage? 2014-06-11 08:28:36 would save 2MB+ from xen package, for those who does not use it 2014-06-11 08:36:51 hmm, it doesn't see shipped with alpine-xen. that comes with pygrub 2014-06-11 08:41:09 ok, got it 2014-06-11 08:42:27 ncopa: suse is shipping Xen, not sure how they package pvgrub 2014-06-11 08:44:28 royger: does it make sense to you to ship pvgrub and possibly hvmloader and py-grub in separate packages? 2014-06-11 08:44:46 it it normal that you use a mix of them or do you pick one and stick to it? 2014-06-11 08:48:27 I wouldn't ship hvmloader in a separate package, the rest could make sense 2014-06-11 08:49:22 pvgrub is the one that makes more sense, pygrub not sure, the only way I could see that as an advantage is because of security, pygrub is not as secure as pvgrub, so it could make sense to ship it in a separate package 2014-06-11 08:49:46 and advertise ppl to use pvgrub when possible 2014-06-11 08:55:11 ok 2014-06-11 08:55:35 yes, if we ship pvgrub and pygrub in separate packages, then users would have to pick one 2014-06-11 08:55:55 is spice-vdagent in one of the spice pkgs available? 2014-06-11 08:56:51 also i packaged w3m earlier today, available in my aports clone on github 2014-06-11 08:57:59 WwQCn4HqahlD: seems like we dont ship spice-vdagent 2014-06-11 08:58:04 we should package it 2014-06-11 08:58:32 we have the xf86-video-qxl package though 2014-06-11 08:59:31 awesome. i'll give it a shot later. 2014-06-11 09:03:01 also another q: is there docs somewhere how to slim down an alpinelinux install? kernel-firmware+imgs take like almost 200M. or is there a seperate kernel for VMs without all the modules packaged? 2014-06-11 09:03:01 ncopa: no, you should be able to have both packages, just like now 2014-06-11 09:05:54 yes you could install both if you wanted 2014-06-11 09:08:24 WwQCn4HqahlD: i just pushed spice-vdagent 2014-06-11 09:08:31 wow thx!!!! 2014-06-11 09:08:42 you'll have to test it from edge/testing repo 2014-06-11 09:08:52 i'll do so! thanks a lot! 2014-06-11 09:09:06 if you want it backported to v3.0 then please create a ticket and set target 3.0.1 2014-06-11 09:09:20 no, i like testing. thx. 2014-06-11 09:09:51 please let me know if it works, so i can move it to 'main' repo 2014-06-11 09:09:56 and it will be included in v3.1 2014-06-11 09:10:28 ok! i will deploy and test it later today! 2014-06-11 09:13:30 ncopa: do we support backports for >2.7 now? 2014-06-11 09:13:47 on request we do 2014-06-11 09:13:57 as long as they dont break anything 2014-06-11 09:14:11 i dont thik we will add any backports/ repo for v3.0 2014-06-11 09:14:33 we just cherry-pick to v3.0 if it can be done non-intrusive 2014-06-11 09:14:40 and only on request 2014-06-11 09:15:19 its basically so we avoid giving support to users that mix edge repo with stable branch when they should run stable only 2014-06-11 09:19:49 also i guess i'll package ksh 93+ i feel so naked without it. 2014-06-11 10:16:34 first error on loading a vm with pv-grub is libxl: cannot execute /usr/lib/xen/boot/pv-grub-x86_64.gz: Permission denied 2014-06-11 10:16:52 it's missing the execute bit 2014-06-11 10:17:36 although I find that wierd 2014-06-11 10:28:33 ScrumpyJack: how are yoy trying to use it? 2014-06-11 10:29:18 ScrumpyJack: it should be specified in the "kernel" param, not the bootloader param 2014-06-11 10:29:39 ScrumpyJack: because it runs inside the VM, as opposed to pygrub which runs on Dom0 2014-06-11 10:29:47 yeah, i noticed that 2014-06-11 10:30:17 i'm getting a grub menu. just tweaking now. I'll take notes 2014-06-11 10:30:30 a grub shell, not a grub menu 2014-06-11 10:30:34 ScrumpyJack: see if something is missing from http://wiki.xen.org/wiki/PvGrub 2014-06-11 10:31:20 i've been looking at that, and others 2014-06-11 11:17:06 ok, got pv-grub working with 2.7.8. I'll try booting 3.0.0 and document 2014-06-11 12:51:07 to make things easier to use pvgrub, can we put a grub conf file in the distro? it takes away a lot of hassle 2014-06-11 12:53:06 http://paste2.org/pxEp7FIk 2014-06-11 12:54:38 after the initial boot off cdrom/iso image when creating a pv guest, and installing 'sys' onto your partition, you must reboot to have the file systems mounted properly. 2014-06-11 13:05:33 to reboot, can't use pvgrub straight away, you must boot the alpine pv guest some other way, (pygrub in alpine-xen doesn't work for me) and then create a grub menu file somewhere 2014-06-11 13:06:17 then you can halt again and use pvgrub. that's one halt too many :) 2014-06-11 13:13:34 ncopa: how did you install alpine on a UEFI system? booted another distro and used apk-static for installation? 2014-06-11 13:24:48 uggedal: i booted usb in "legacy" mode 2014-06-11 13:25:23 I reused an existing windows partition 2014-06-11 13:26:02 first week i mount /boot on the usb and used that as the "dual" boot while figuring out how efi works 2014-06-11 13:26:17 whats annoying is that you cannot reach efivars unless you boot in efimode 2014-06-11 13:26:48 so even if i could mount and copy in gummiboot files to the fat efi boot partition 2014-06-11 13:26:57 i could not replace the boot manager from linux 2014-06-11 13:27:09 so i ended up run some magic command from windows 2014-06-11 13:29:19 i think i used bcdboot 2014-06-11 13:29:44 hm 2014-06-11 13:29:46 maybe not 2014-06-11 13:30:01 bcdedit 2014-06-11 13:30:58 ah, I think I'll boot archlinux iso and use apk-static then :) 2014-06-11 13:32:02 i suppose we need document how to create a boot usb for efi 2014-06-11 13:32:30 the archlinux iso is isohybrid? 2014-06-11 13:32:38 yeah 2014-06-11 13:33:00 but I'm booting it as an cd-rom now from virtualbox to test things 2014-06-11 13:33:15 you can boot cdrom in efi mode too? 2014-06-11 13:33:19 i didnt know that :) 2014-06-11 13:34:04 yeah, but my new machine will not have a cdrom. but archlinux boots fine from usb 2014-06-11 13:34:32 even on my macbookpro which has a broken EFI implementation 2014-06-11 13:35:18 are you booting straight from gummitboot to the kernel (EFISTUB)? 2014-06-11 13:35:35 gummiboot* 2014-06-11 13:35:38 yes i think so 2014-06-11 13:35:49 i'll check my gummiboot.conf 2014-06-11 13:38:59 yes 2014-06-11 13:39:57 hm 2014-06-11 13:40:07 i tihnk we need a trigger script for gummiboot 2014-06-11 13:40:25 to generate the loader/entries/* 2014-06-11 13:41:18 $ cat /boot/loader/loader.conf 2014-06-11 13:41:18 timeout 5 2014-06-11 13:41:18 default grsec 2014-06-11 13:41:30 $ ls /boot/loader/entries/ 2014-06-11 13:41:30 grsec.conf vanilla.conf 2014-06-11 13:41:47 $ cat /boot/loader/entries/grsec.conf 2014-06-11 13:41:47 title Alpine Linux Grsecurity Kernel 2014-06-11 13:41:47 efi /vmlinuz-grsec 2014-06-11 13:41:47 options root=/dev/vg0/lv_root modules=sd-mod,ext4 quiet i8042.notimeout initrd=initramfs-grsec 2014-06-11 13:42:09 $ cat /boot/loader/entries/vanilla.conf 2014-06-11 13:42:10 title Alpine Linux Vanilla Kernel 2014-06-11 13:42:10 efi /vmlinuz 2014-06-11 13:42:10 options root=/dev/vg0/lv_root modules=sd-mod,ext4 quiet i8042.notimeout initrd=initramfs-vanilla 2014-06-11 13:42:56 gummiboot is also in 'testing' repo 2014-06-11 13:59:34 running setup-alpine in 3.0.0: modprobe: can't change directory to '3.10.40-0-grsec': No such file or directory 2014-06-11 14:18:29 ignore that 2014-06-11 15:08:30 fixed things in http://wiki.alpinelinux.org/wiki/Create_Alpine_Linux_PV_DomU and updated for use with pvgrub 2014-06-11 16:03:22 ping tdtrask 2014-06-11 17:16:35 i hope it's ok to just go update the wiki. there is a (old) list on bugtracker i see. #589. 2014-06-11 18:14:45 ncopa: well, i dont have much opinion 2014-06-12 00:39:01 alacerda: in case you're reading the logs, pong 2014-06-12 01:03:53 anyone using openldap on musl edge 2014-06-12 01:03:54 ? 2014-06-12 07:15:25 Morning 2014-06-12 11:09:38 alpine linux 3 review http://news.softpedia.com/news/Alpine-Linux-3-0-0-Is-an-OS-for-Terminal-Aficionados-445364.shtml 2014-06-12 11:33:13 cool 2014-06-12 11:33:29 the review sounds like a copy from release notes 2014-06-12 11:33:37 but they did the job of making a screenshot 2014-06-12 11:33:47 so they have actually installed it likely 2014-06-12 11:33:58 ah 2014-06-12 11:34:11 i also notice that xfce looks ugly by default 2014-06-12 11:34:22 i think it gets the theme wrong 2014-06-12 11:34:25 i dunno why 2014-06-12 11:34:33 that is something i'd like to fix 2014-06-12 11:34:39 what i want is: setup-xfce-desktop 2014-06-12 11:34:42 and boom... 2014-06-12 11:34:49 you have a nice looking default desktop 2014-06-12 11:37:05 or maybe: setup-desktop xfce4 2014-06-12 11:45:21 ncopa: what is the current status of nfs on 64bit? 2014-06-12 11:46:30 im thinking of upgrading my home server, and if nfs still isnt working, probably i can run lxc for it. 2014-06-12 11:48:09 i dont know if nfs runs inside lxc 2014-06-12 11:48:20 as nfs is a kernel thing 2014-06-12 11:48:26 why wouldnt it? 2014-06-12 11:48:40 because its a kernel feature? 2014-06-12 11:48:46 i can pass all that is needed i guess? 2014-06-12 11:48:57 and load the kernel modules on the host 2014-06-12 11:48:59 might work 2014-06-12 11:49:07 i run kvm inside lxc 2014-06-12 11:49:14 and kvm is a kernel thing :) 2014-06-12 11:49:25 you are right, it can get hacky 2014-06-12 11:49:43 specially since it deals with network namespaces 2014-06-12 11:49:51 maybe ask in #lxcontainers? 2014-06-12 11:50:12 so i guess this answers my first question? 2014-06-12 11:50:15 :) 2014-06-12 11:50:31 http://comments.gmane.org/gmane.linux.kernel.containers.lxc.general/5791 2014-06-12 11:50:32 yes 2014-06-12 11:50:42 no progress wrt nfs on 64bit :-/ 2014-06-12 11:50:53 but i will need to fix it at some point 2014-06-12 11:51:09 its nasty 2014-06-12 11:51:17 because it just hangs 2014-06-12 11:51:23 and there are N components 2014-06-12 11:51:26 and I have no 2014-06-12 11:51:32 idea where it gets stuck 2014-06-12 11:51:36 bummer with nfs on lxc 2014-06-12 11:52:26 ill just wait for your magic to happen. 2014-06-12 11:52:40 might be i'll have to wait for fabled's magic to happen :-/ 2014-06-12 11:52:50 i was about to ask that :) 2014-06-12 11:53:18 fabled can fix anythning 2014-06-12 11:53:33 the impossible just takes a little more time :) 2014-06-12 11:53:51 i bet he can evne fix my spelling mistakes :p 2014-06-12 11:54:42 :) 2014-06-12 11:58:07 ncopa: i have a working freeswitch build for x86_64 2014-06-12 11:58:18 but you added some musl patches 2014-06-12 11:58:36 i guess those didnt introduce build breakage? 2014-06-12 11:59:23 did you fix the bundled 3rdparty libs? 2014-06-12 11:59:49 what should i fix? 2014-06-12 11:59:58 i just removed some part 2014-06-12 12:00:04 ok let me ask this way 2014-06-12 12:00:19 did you make it build with the config we shipped with v2.7? 2014-06-12 12:00:37 yes 2014-06-12 12:00:47 it was breaking on libshout-2.2.2 2014-06-12 12:00:57 aha 2014-06-12 12:00:58 DSO issues 2014-06-12 12:01:07 but it was in the examples dir 2014-06-12 12:01:17 oh.. 2014-06-12 12:01:18 i just removed that from makefile 2014-06-12 12:01:26 rebased your patches 2014-06-12 12:01:38 and it built then? 2014-06-12 12:01:42 for lua and textrels 2014-06-12 12:01:46 yeah 2014-06-12 12:01:51 nice! 2014-06-12 12:01:54 very very nice 2014-06-12 12:02:11 of course on latest version 2014-06-12 12:02:24 im not saying it works 2014-06-12 12:02:27 oh.. whats the latest version.... 2014-06-12 12:02:28 but it builds 2014-06-12 12:02:39 1.4.6 2014-06-12 12:02:41 i figured out that "latest version" is a non-trivial question :) 2014-06-12 12:03:00 oh 2014-06-12 12:03:14 they made 1.4.x releases! 2014-06-12 12:03:22 they ship a tarball for 1.4.6? 2014-06-12 12:03:24 thats just a few minor bumps to our 1.4.0 2014-06-12 12:03:35 yes 2014-06-12 12:03:36 \o/ 2014-06-12 12:03:49 wonderful! finally! 2014-06-12 12:03:56 i spoke to a guy yerterday 2014-06-12 12:04:16 1.4.2 and earlier is a mess wrt release engineering 2014-06-12 12:04:38 i spoke to a guy on #freeswitch last month or so 2014-06-12 12:04:45 i think he tried to be helpful 2014-06-12 12:05:09 but it felt like he had ad hd or was on speed or something 2014-06-12 12:05:22 but i think he really tried to help 2014-06-12 12:05:45 ok 2014-06-12 12:05:48 but thats good new 2014-06-12 12:05:55 it means we can ship 1.4.6 2014-06-12 12:06:01 he said there was some intiative in trunk to move their libs to std libs 2014-06-12 12:06:09 i noticed that 2014-06-12 12:06:26 i think they finally realized the value of shared libs 2014-06-12 12:06:41 one thing about your patch 2014-06-12 12:06:45 ok? 2014-06-12 12:06:50 which of them? 2014-06-12 12:07:07 - char chars[TZ_BIGGEST(TZ_MAX_CHARS + 1, (2*(TZNAME_MAX + 1)))]; 2014-06-12 12:07:07 + char chars[TZ_BIGGEST(TZ_MAX_CHARS + 1, (2*(SPANDSP_TZNAME_MAX + 1)))]; 2014-06-12 12:08:00 its in http://git.alpinelinux.org/cgit/aports/tree/main/freeswitch/musl-fixes.patch 2014-06-12 12:08:12 is it not needed? 2014-06-12 12:08:25 not that i noticed 2014-06-12 12:08:49 or maybe they moved spandsp out of there tree 2014-06-12 12:08:59 might be something was fixed in either musl or spandsp 2014-06-12 12:09:18 i think spandsp isnt shared lib 2014-06-12 12:09:46 i think that change was due to a name clash 2014-06-12 12:09:56 -#define TZNAME_MAX 255 2014-06-12 12:09:56 +#define SPANDSP_TZNAME_MAX 255 2014-06-12 12:10:09 looks like i got error that TZNAME_MAX was already defined 2014-06-12 12:10:17 might be they disabled -Werror also 2014-06-12 12:10:46 ill cleanup the build and push 2014-06-12 12:10:57 wonderful! 2014-06-12 12:11:04 lets see what other archs think about it 2014-06-12 12:11:20 heh 2014-06-12 12:11:24 poor arm 2014-06-12 12:11:28 will likely build a week ro so :) 2014-06-12 12:11:30 or* 2014-06-12 12:15:59 yeah, the configure part is madness... 2014-06-12 12:17:00 configure part takes longer time than the actual building 2014-06-12 12:19:10 ncopa: until it catch up with commits after freeswitch, it'll be 3 weeks :D 2014-06-12 12:19:24 :) 2014-06-12 13:14:52 \o/ 2014-06-12 13:21:17 clandmeter: thanks!!! 2014-06-12 13:33:29 nice fix to disable examples 2014-06-12 13:33:34 should reduce build time too 2014-06-12 13:37:19 i'm impressed with armhf. how is that coming on? 2014-06-12 13:46:47 clandmeter: good job! 2014-06-12 13:47:00 ScrumpyJack: armhf is progressing slowly 2014-06-12 13:47:51 :) 2014-06-12 13:55:22 task warrior (task) depends on libgcc in aports. does that mean that it doesn't compile with musl? 2014-06-12 14:04:45 ScrumpyJack: no. libgcc comes from gcc 2014-06-12 14:05:12 and task warrior seems to work here 2014-06-12 14:28:18 works for me too, just wondered why it needs libgcc 2014-06-12 16:48:29 Would it be possible to build VirtualBox on Apline 3.0? 2014-06-12 16:48:42 Alpine 3.0 2014-06-12 17:35:58 ping ncopa 2014-06-12 18:04:20 krjn: possibly yes, but it's not easy 2014-06-12 20:04:50 alacerda: ping 2014-06-12 20:15:47 ncopa: ayy, silly me 2014-06-12 20:16:03 ncopa: can you remind me how to create a git repo on git.al.o? 2014-06-12 20:16:45 I remember that I need to create bare repo, but that's all 2014-06-12 20:16:52 nothing to get it actually working on listed in cgit… 2014-06-12 20:18:37 s/on/or/ 2014-06-12 21:49:11 barthalion: yeah, the list of dependencies is ridiculous 2014-06-13 00:05:12 just got home 2014-06-13 00:05:27 I'll reply to the openldap bug tonight maybe tomorrow ncopa 2014-06-13 00:25:25 holaa 2014-06-13 01:52:18 barthalion: http://wiki.alpinelinux.org/wiki/Development_using_git:Cgit 2014-06-13 06:58:59 morning 2014-06-13 10:13:51 tdtrask: thank you 2014-06-14 09:40:07 morning 2014-06-14 09:40:18 I'm building my first alpine package (yay!) 2014-06-14 09:40:27 but the docs are a bit sparse 2014-06-14 09:41:33 does someone have an up to date summary? 2014-06-14 09:42:14 so before i start ... 2014-06-14 09:42:34 better ask your questions here 2014-06-14 09:42:40 there is APKBUILD reference on wiki 2014-06-14 09:44:25 abuild doesn't actually compile the package does it? (like rpmbuild or debuild?) 2014-06-14 09:44:35 no, it's kind of wrapper 2014-06-14 09:44:53 you basically write set of steps needed to build a package, like a shell script 2014-06-14 09:45:12 + required (make)depends and other metadata 2014-06-14 09:45:43 ok 2014-06-14 09:46:31 i'm looking at this http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2014-06-14 09:47:08 and just found this http://wiki.alpinelinux.org/wiki/APKBUILD_Reference 2014-06-14 09:47:11 thanks for the pointer 2014-06-14 09:47:21 np 2014-06-14 10:02:30 hi all 2014-06-14 10:03:03 just giving a look at http://bugs.alpinelinux.org/issues/3006 2014-06-14 10:03:22 i've almost done all the packages openca-* related 2014-06-14 10:03:52 but libpki returns a strange error when linking (even if the pacakge compiles) 2014-06-14 10:03:57 *package 2014-06-14 10:04:19 multiple -rpath found on /usr/lib 2014-06-14 10:04:31 what does it mena? 2014-06-14 10:04:34 *mean? 2014-06-14 10:10:52 and, then: >>> ERROR: libpki*: libpki.so.85: path not found 2014-06-14 12:39:32 fcolista: what happens if you remove rpathes with chrpath or patchelf? 2014-06-14 12:40:05 how can i do that barthalion ? 2014-06-14 12:40:10 patching makefile? 2014-06-14 12:40:31 or configure script? 2014-06-14 12:40:45 Makefile.in:rpath = @rpath@ 2014-06-14 12:41:38 fcolista: well, fixing it in Makefile/configure would be the best 2014-06-14 12:41:46 but it's usually not simple 2014-06-14 12:41:52 at least for minors like me :) 2014-06-14 12:41:55 http://directory.fsf.org/wiki/Chrpath 2014-06-14 12:41:59 should be in main 2014-06-14 12:42:55 i don't know what rpath is supposed to do. If is only matter of changing rpath command with chrpath it's ok 2014-06-14 12:43:21 well, rpath is actually not a command 2014-06-14 12:43:37 but you understood what i meant 2014-06-14 12:44:25 ah 2014-06-14 12:44:26 ok 2014-06-14 12:44:38 umh.. 2014-06-14 12:45:10 usage looks like this: chrpath -d "$pkgdir"/usr/lib/libpki.so.85 2014-06-14 12:45:42 do we have chrpath ? 2014-06-14 12:46:03 yes we do 2014-06-16 07:39:48 Morning 2014-06-17 08:48:18 Morning 2014-06-17 09:24:22 is CVE-2014-4014 fix included in grsecurity-3.0-3.14.6-201406101411.patch? 2014-06-17 09:24:32 seems like fix is inclued in 3.14.8 upstream 2014-06-17 09:25:37 wrong chan... 2014-06-17 09:59:58 ncopa: fsck.vfat segfaults 2014-06-17 10:00:22 care to file a bug? 2014-06-17 10:00:34 i first need to fix my system 2014-06-17 10:00:40 i think there was something with syslinux/mtools 2014-06-17 10:00:51 and i think there was a fix for mtools iirc 2014-06-17 10:00:52 upgrade was not so smooth, to say the least. 2014-06-17 10:00:56 :-( 2014-06-17 10:02:45 damn i cant umount usb, nor mount it as rw 2014-06-17 10:04:15 looks like something is automagically mounting it back as ro 2014-06-17 10:12:19 ncopa: i think we should really have openssh in alpine-mini 2014-06-17 10:15:09 for sure when upgrading, it would be nice to have access to headless. 2014-06-17 10:49:44 a small silent question: 2014-06-17 10:49:51 any news regarding nfs? 2014-06-17 10:52:38 StarWarsFan: sorry not yet :-( 2014-06-17 10:52:46 i am stuck there really 2014-06-17 10:59:37 hm, not good 2014-06-17 10:59:58 seems to be a very strange issue 2014-06-17 11:04:18 StarWarsFan: send flowers to fabled and hope he has some free time. 2014-06-17 11:05:06 i have already spent various hours on the issue 2014-06-17 11:06:50 for fabled: 2014-06-17 11:06:52 _ 2014-06-17 11:06:53 _(_)_ wWWWw _ 2014-06-17 11:06:53 @@@@ (_)@(_) vVVVv _ @@@@ (___) _(_)_ 2014-06-17 11:06:53 @@()@@ wWWWw (_)\ (___) _(_)_ @@()@@ Y (_)@(_) 2014-06-17 11:06:53 @@@@ (___) `|/ Y (_)@(_) @@@@ \|/ (_)\ 2014-06-17 11:06:55 / Y \| \|/ /(_) \| |/ | 2014-06-17 11:06:57 \ | \ |/ | / \ | / \|/ |/ \| \|/ 2014-06-17 11:07:00 \\|// \\|/// \\\|//\\\|/// \|/// \\\|// \\|// \\\|// 2014-06-17 11:07:02 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 2014-06-17 11:07:04 ;-) 2014-06-17 11:07:13 (as requested...) 2014-06-17 11:07:48 lol 2014-06-17 11:09:21 maybe we are lucky and dalias needs it on his fresh alpine install. 2014-06-17 11:14:36 ncopa: you may want to take a look at bash patch used in Arch now 2014-06-17 11:14:56 ncopa: I've emailed bash maintainer about the problem because I haven't found any trace of reporting it by researcher 2014-06-17 11:15:36 is the bash patch i just pushed to alpline bad? 2014-06-17 11:16:12 basically Chet said that bash shouldn't exit when setuid fails, just emit a warning 2014-06-17 11:16:58 but EXIT_ON_SETUID_FAILURE can be defined at build time if you really want it to exit 2014-06-17 11:17:13 ok, going back to work :) 2014-06-17 11:19:02 ugh, wtf is wrong with my nick 2014-06-17 11:24:21 ncopa: did you ever see this when creating a container? 2014-06-17 11:24:22 (1/15) Installing musl (1.1.2-r1) 2014-06-17 11:24:22 Executing musl-1.1.2-r1.post-install 2014-06-17 11:24:22 ERROR: musl-1.1.2-r1.post-install: script exited with error 1 2014-06-17 11:25:21 ah thats timezone 2014-06-17 11:25:29 didnt you change something recently 2014-06-17 11:25:41 yes 2014-06-17 11:26:16 i wonder why it exits with error 2014-06-17 11:27:20 i think the logic is going wrong 2014-06-17 11:27:27 if ! [ -e /etc/localtime ] && [ -e /etc/zoneinfo/localtime ]; then 2014-06-17 11:27:43 i guess so 2014-06-17 11:27:45 i think it ends up as: if ! ([ -e /etc/localtime ] && [ -e /etc/zoneinfo/localtime ]); then 2014-06-17 11:28:44 if ! [ -e /etc/localtime -a -e /etc/zoneinfo/localtime ]; then 2014-06-17 11:28:51 i think we need to change it to that 2014-06-17 11:29:43 no 2014-06-17 11:30:02 if [ ! -e /etc/localtime ] && [ -e /etc/zoneinfo/localtime ]; then 2014-06-17 11:30:33 yes 2014-06-17 11:35:54 no 2014-06-17 11:36:07 problem is circular dep 2014-06-17 11:36:44 busybox depends on musl 2014-06-17 11:37:05 musl's trigger script depends on /bin/sh provided from busybox 2014-06-17 11:37:12 so problem is that /bin/sh is not there 2014-06-17 11:37:18 ah ok 2014-06-17 11:37:30 i suppose we could write the trigger in C 2014-06-17 11:41:16 hmm zoneminder depends on apache2 2014-06-17 14:06:26 anyone running openstack on alpine? 2014-06-17 15:44:55 ncopa no, but i could try. why the question? 2014-06-18 07:52:11 morning 2014-06-18 07:52:29 apk add lxc 2014-06-18 07:52:34 lxc-ls 2014-06-18 07:52:53 Traceback (most recent call last): 2014-06-18 07:52:53 File "/usr/bin/lxc-ls", line 31, in 2014-06-18 07:52:53 import lxc 2014-06-18 07:55:08 erm, apologies. so sorry, I was in arch linux! :) 2014-06-18 08:12:54 :D 2014-06-18 08:18:29 I'm desperate to get alpine LXCs working so that I can start building aports 2014-06-18 08:19:12 I can build and tear down alpine containers on a hosted arch VM 2014-06-18 08:19:53 but i want to try getting NAT to work so that I can get network from within the alpine container 2014-06-18 08:20:05 (I really don't want to bridge) 2014-06-18 08:20:45 anyone seem some decent docs? 2014-06-18 08:27:06 are you using netctl over wifi? 2014-06-18 08:27:40 "netctl lxc" gives reasonable amount of howtos, but only over a cable 2014-06-18 08:28:30 you need to play with ebtables to deceive router 2014-06-18 09:04:20 i'm not using anything at the moment, all my LXC network.type = empty. I'm thinking about the best way to give my alpine LXC network access without bridging. one ethernet port of a hosted server, no wifi involved 2014-06-18 09:06:25 I'm just learning about the net namespace. 2014-06-18 09:10:11 ScrumpyJack: maybe macvlan is an option? 2014-06-18 09:34:32 that's what I was hoping to use somehow. but still figuring out how. creating some macvlan NICs in a private subnet on the host and giving one to the container namespace isn't enough. the macvlan NIC swallowed up by the container can't talk to the other macvlan NICs in the host 2014-06-18 09:57:55 ScrumpyJack: to my understanding you need create a macvlan on the host aswell and use that 2014-06-18 10:00:53 that's what I'm doing. macvlan device in the host, and type phy in the container, where the type phy is pointed to the macvlan interface created on the host 2014-06-18 10:01:48 files from v2.5.4-309-gadd245d uploaded 2014-06-18 10:02:04 by the way: the LXC script that checks for alpine version still say "Determining the latest release... v2.7" 2014-06-18 10:07:22 Using static apk from http://wiki.alpinelinux.org/cgi-bin/dl.cgi/v2.7/main/x86_64 2014-06-18 10:25:47 hello 2014-06-18 10:26:56 question: how hard would it be for someone who's not really a developer to cross compile alpine for arm? :) 2014-06-18 10:27:48 the famous question: why would you do it? we have arm bootstrapped and working already 2014-06-18 10:27:54 there are some test images somewhere 2014-06-18 10:29:42 hm, to make it short: i got the feeling that arm isnt really supported, so i wanted to try and compile current stable for it 2014-06-18 10:30:06 not being an expert, i have no idea if this is really hard or fairly simple 2014-06-18 10:32:30 it's probably somewhere in between, however you've got wrong feeling :p 2014-06-18 10:32:58 :D 2014-06-18 10:34:56 thing is, i actually wanted to contribute a few years back, but due to time constraints i never got to it - now i have a bit of a bad feeling about it 2014-06-18 10:35:04 I don't think I've bookmarked link to test images 2014-06-18 10:35:52 ncopa, fabled, clandmeter ↑ 2014-06-18 10:36:22 i'd want to help you guys, but i'm not a programmer; i'm actually a network guy... if porting to arm involves just fiddling with configuration and getting it to boot after it's compiled, then i can see what i can do 2014-06-18 10:38:16 i got one image from the mailing list :) 2014-06-18 10:38:53 thanks 2014-06-18 11:00:26 well crosscompiling is non-trivial 2014-06-18 11:03:36 hey, waydot. you are a network guy, do you anything about linux containers and networking? I'm trying to build a linux container with IP, but no bridging involved. 2014-06-18 11:35:22 ScrumpyJack: by container you mean for example KVM or is it something specific? 2014-06-18 11:36:21 to me it seems as if you need to use bridging for kvm, but i havent given it a lot of time 2014-06-18 11:52:44 lxc 2014-06-18 11:53:00 https://linuxcontainers.org/ 2014-06-18 12:10:34 was there any discussion of moving /[s]bin to /usr/[s]in ? 2014-06-18 12:10:36 http://www.freedesktop.org/wiki/Software/systemd/TheCaseForTheUsrMerge/ 2014-06-18 12:11:12 or still early, as only few distro have implemented it 2014-06-18 12:12:29 /usr/[s]bin* 2014-06-18 12:16:19 oops 2014-06-18 12:19:47 we just follow upstream 2014-06-18 12:19:56 lxc 2014-06-18 12:22:12 interesting read :) (especially the mail from rob landley) 2014-06-18 12:22:48 ScrumpyJack: sorry, i havent tried those 2014-06-18 12:23:58 on BSD separate /usr works nicely, though :) 2014-06-18 12:24:08 I have a host, with one eth. I create a macvlan device in the host, and put it on a subnet. I create a LXC with a network.type macvlan in bridge mode with my host macvlan device and I put it on the same subnet as my host macvlan device. nothing :( 2014-06-18 12:26:54 nothing on the host or nothing on the network ? 2014-06-18 12:27:27 i havent used macvlan before, but the way they describe it here http://www.pocketnix.org/posts/Linux%20Networking:%20MAC%20VLANs%20and%20Virtual%20Ethernets you'll need bridging for LXC and host to communicate 2014-06-18 12:45:44 waydot: no tcp between the host and the container 2014-06-18 12:46:26 ScrumpyJack: that requires bridge mode 2014-06-18 12:47:02 yup, done that 2014-06-18 12:47:20 ScrumpyJack: oh, and it still doesnt work? 2014-06-18 12:47:46 I create a container with a network.type macvlan in bridge mode with my host macvlan device 2014-06-18 12:48:07 i'm certain that i'm doing something wrong. 2014-06-18 12:49:01 i want to avoid bridging my eth device on the host, so i create a macvlan on the host and bridge that to the macvlan in the container 2014-06-18 12:49:36 ScrumpyJack: do you have a private net between host and container 2014-06-18 12:49:49 or is your containers supposed to be in same subnet as the host? 2014-06-18 12:50:54 i have some boxes with a private net for the containers 2014-06-18 12:51:09 to let those talk with the host i have added a dummy interface on the host 2014-06-18 12:51:30 and created a bridge 2014-06-18 12:51:34 i put the macvlan created on the host in 192.168 and the macvlan in bridge mode in the container on the same network. they can't ping 2014-06-18 12:52:33 the eth device on the host is a public net 2014-06-18 12:52:34 macvlan is to my understanding useful if you want the container talk directly with the switch 2014-06-18 12:52:54 are your containers supposed to be on the same public net? 2014-06-18 12:53:12 yes, but through NAT 2014-06-18 12:53:18 ok 2014-06-18 12:53:35 similar to what i do on our bld[12].alpinelinux.org boxes 2014-06-18 12:53:40 we have limited public ip addresses 2014-06-18 12:53:53 so all containers run in their private 192.168 net 2014-06-18 12:54:12 ah cool. is your public eth device bridged? 2014-06-18 12:54:27 no 2014-06-18 12:54:32 well 2014-06-18 12:54:35 for some of the guests 2014-06-18 12:54:40 or have you created other devices to bridge 2014-06-18 12:55:01 auto lxbr0 2014-06-18 12:55:01 iface lxbr0 inet static 2014-06-18 12:55:01 bridge-ports dummy0 bond1 2014-06-18 12:55:01 bridge-stp off 2014-06-18 12:55:01 address 192.168.8.1 2014-06-18 12:55:02 netmask 255.255.255.0 2014-06-18 12:55:32 bond1 is crossover cable to the other build server 2014-06-18 12:55:38 yeah, so your public interface is br0 ? 2014-06-18 12:55:54 correct 2014-06-18 12:56:00 i can't do that 2014-06-18 12:56:21 i need to create other devices (macvlan) 2014-06-18 12:56:34 but the only reason i do that is that i have a few containers that i give a public ip 2014-06-18 12:56:51 if all containers only have private ip, then the br0 is not needed 2014-06-18 12:56:53 same :) i only have one public ip 2014-06-18 12:57:11 you could drop the bond1 then 2014-06-18 12:57:24 and only one host that has that private net? 2014-06-18 12:57:46 i create a dummy interface on the host for the private net 2014-06-18 12:57:51 modprobe dummy 2014-06-18 12:58:10 aha! 2014-06-18 12:58:20 # cat /etc/lxc/default.conf 2014-06-18 12:58:20 lxc.network.type = veth 2014-06-18 12:58:20 lxc.network.link = lxbr0 2014-06-18 12:58:20 that might be what I'm doing wrong 2014-06-18 12:58:39 I'm not creating a dummy0, i'm creating a macvlan 2014-06-18 12:58:41 i think you can make it work with macvlan too, but i don't think that is what you want 2014-06-18 12:58:54 ok, lemme test 2014-06-18 12:59:31 i think you want macvlan if you want the containers in the public net 2014-06-18 12:59:53 or if you for some reason dont want the containers be able to talk with the host at all 2014-06-18 13:06:46 oh man, arch linux is such a pain. there was nothing wrong with init 2014-06-18 13:23:14 ok, so by bridge device has ip 192.168.1 and it is composed by a dummy interface 2014-06-18 13:24:27 my lxc interface type is veth and it is linked to my bridged device (br0) 2014-06-18 13:24:41 and has an IP on the same network. 2014-06-18 13:24:51 container and host can't ping :( 2014-06-18 13:26:05 i notice that this has create another device in the host vethKEWCD4 2014-06-18 13:27:58 what is that? a pipe? 2014-06-18 13:33:45 yup 2014-06-18 13:34:06 brctl show br0 2014-06-18 13:35:04 verify that br0 has both dummy0 and the veth interface 2014-06-18 13:35:10 ip link br0 2014-06-18 13:35:18 verify that link is up 2014-06-18 14:12:55 my link was down 2014-06-18 14:12:58 silly me 2014-06-18 14:18:03 cool, NAT works now. 2014-06-18 14:18:42 the setup-alpine is all over the place in a container. perhaps there should be a container specific setup :) 2014-06-18 14:19:43 i actually thought about setup-alpine in container the other day 2014-06-18 14:20:09 i should not be too hard to make it do the right thing in a container 2014-06-18 14:20:23 i just have not had the time to actually do it 2014-06-18 14:20:44 ScrumpyJack: do you mind file an issue on bugs.a.o for setup-alpine in a container? 2014-06-18 14:20:49 so i dont forget... 2014-06-18 15:27:45 ncopa: certainly 2014-06-18 17:44:01 . 2014-06-18 17:56:35 added bits to the alpine LXC container wiki page about bridging without interfering 2014-06-19 07:20:21 morning 2014-06-19 07:26:37 ncopa: you about? 2014-06-19 07:26:59 hi 2014-06-19 07:30:10 hi. time to make some packages. just a few questions as i go on. i have pulled the aports. where should my package directory be? 2014-06-19 07:30:30 ScrumpyJack: initially in testing/ 2014-06-19 07:31:15 once the package is built by the autobuilder, and you have verified that the autobuilders version is ok 2014-06-19 07:31:17 so i run newapkbuild -c mypackage in testing? 2014-06-19 07:31:23 then we can move it to main 2014-06-19 07:31:24 yes 2014-06-19 07:31:31 ncopa: nice, i was about to fix that fpm script. 2014-06-19 07:31:39 cool, thanks 2014-06-19 07:31:49 does depend need to be an extra cmd? 2014-06-19 07:32:01 clandmeter: np :) i was looking over the bugs for 2.7.9 2014-06-19 07:32:17 clandmeter: i think not actually. i suspect gentoo devs did that wrong 2014-06-19 07:32:24 i guess so 2014-06-19 07:32:35 i just copied it from gentoo and cleaned it up slightly 2014-06-19 07:32:52 now you an run php-fpm depend (which makes no sence) :) 2014-06-19 07:32:59 :) 2014-06-19 07:33:01 true 2014-06-19 07:33:08 i should have removed that 2014-06-19 07:33:24 would it be a problem if i build outside the aports tree and more it into testing when finished? I can't see that it would but you never know 2014-06-19 07:33:38 no, it should be no problem 2014-06-19 07:33:49 i actually have an aports/wip dir 2014-06-19 07:33:55 where i add all new stuff 2014-06-19 07:34:07 and a script that copies it to testing/ and do git commit for me 2014-06-19 07:45:24 cool 2014-06-19 07:45:51 what about man pages. i guess you'd rather they were in a package-doc package? 2014-06-19 07:51:34 ScrumpyJack: yes 2014-06-19 07:51:43 just add $pkgname-doc to subpackages 2014-06-19 07:53:57 yes, i got that, but i'm trying to figure out how that works. 2014-06-19 07:54:25 when i make the package, it fails with can't cd to $pkgname-doc 2014-06-19 07:54:38 so i need to make a new package just for the docs? 2014-06-19 07:54:57 let me read more on that 2014-06-19 07:56:32 http://wiki.alpinelinux.org/wiki/APKBUILD_examples:Subpackages not much in there 2014-06-19 07:57:19 tehre is a function in abuild that will find all files that it thinks belongs to the -doc subpackage 2014-06-19 07:57:39 if it does not find any files it will not create the $pkgname-doc dir and error out with that message 2014-06-19 07:57:40 the doc() function? http://wiki.alpinelinux.org/wiki/APKBUILD_examples:Special_Subpackages_Doc 2014-06-19 07:57:55 run: find pkg/$pkgname 2014-06-19 07:58:19 if there are no /usr/share/{man,doc,info} or similar it will not create the pkg/$pkgname-doc dir 2014-06-19 07:58:41 so manpages are either not created or installed in unexpected location (maybe /usr/local?) 2014-06-19 07:59:43 run find where? 2014-06-19 08:01:22 ScrumpyJack: in your $startdir, where you APKBUILD is 2014-06-19 08:12:53 there is no pkg dir 2014-06-19 08:13:11 only src 2014-06-19 08:29:55 does you package do any make install? 2014-06-19 08:44:45 not yes, i'm just doing a make 2014-06-19 08:45:05 so pkg is the fakeroot? 2014-06-19 10:34:04 >>> sc: Entering fakeroot.. 2014-06-19 10:34:15 where do i define the pkgdir? what is the default? 2014-06-19 10:37:02 make DESTDIR="$pkgdir" install seems to install in / 2014-06-19 11:01:12 you'll have to look at the Makefile 2014-06-19 11:01:21 they might not use DESTDIR 2014-06-19 11:08:44 DESTDIR is usually prefixed in Makefiles no? 2014-06-19 11:09:14 stuff gets installed in $prefix/$DESTDIR 2014-06-19 11:09:48 sounds like someone wrote the makefile reverse of whats common 2014-06-19 11:09:50 where $prefix=/usr and $DESTDIR=$prefix/bin 2014-06-19 11:10:18 try: make install prefix="$pkgdir" DESTDIR=/usr 2014-06-19 11:11:07 I'll play with the Makefile 2014-06-19 11:12:29 actually, that's a silly idea. best to play with APKBUILD :) 2014-06-19 11:13:47 is the makefile available some place? 2014-06-19 11:13:51 can i have a look at it? 2014-06-19 11:13:58 sprunge < Makefile 2014-06-19 11:15:23 I'm sure you have better to do :) 2014-06-19 11:15:55 like can you tell me what makes the pkgdir? when does it get created? 2014-06-19 11:18:40 i am running abuild inside the directory made by newpkgbuild, which holds the APKBUILD file. is that right? 2014-06-19 11:24:11 ignore me 2014-06-19 11:26:34 so pkgdir is pkg/pkgname ? 2014-06-19 11:30:20 http://sprunge.us/GWPJ 2014-06-19 11:30:39 http://sprunge.us/SCje 2014-06-19 11:30:53 http://sprunge.us/bcTI 2014-06-19 11:57:57 ScrumpyJack: abuild creates it 2014-06-19 11:58:08 it doesn't really matter what's the name behind $pkgdir 2014-06-19 12:11:52 barthalion: thank you for your help, i guess I'd better look myself 2014-06-19 12:15:38 EXDIR 2014-06-19 12:15:53 ScrumpyJack: i think you need to drop DESTDIR completely 2014-06-19 12:15:55 they dont implement it 2014-06-19 12:16:21 make install prefix="$pkgdir"/usr will probably work 2014-06-19 12:18:45 i'm not using DESTDIR, i'm using make EXDIR="$pkgdir" install || return 1 2014-06-19 12:20:08 that create /path/pkg/package, but they copies binary into /path/pkg/package/ instead of /path/pkg/package/usr/bin so i need to add that 2014-06-19 12:21:50 you're right, i think make install prefix="$pkgdir"/usr will work 2014-06-19 12:29:15 ok, that creates the right tree, but i'm still getting the No such file or directory with ragard to that 2014-06-19 12:29:36 i'll keep looking 2014-06-19 12:33:56 ScrumpyJack: the directory needs to exist, the makefile uses straight cp for installation, not install -D 2014-06-19 12:34:12 so, mkdir -p "$pkgdir"/usr/bin etc 2014-06-19 12:34:28 and make prefix="$pkgdir"/usr install 2014-06-19 12:40:27 hmm, i'd have thought abuild would make those. I did an abuild on cmus/APKBUILD in aports and abuild created $pkgdir 2014-06-19 12:43:05 pkgdir="$pkgbasedir/$pkgname", ok 2014-06-19 12:48:53 got it 2014-06-19 13:07:24 i indeed needed to create the dirs 2014-06-19 13:13:15 ok, my first package! yay! 2014-06-19 13:13:42 congrats! :-) 2014-06-19 13:13:47 it's in aports/testing 2014-06-19 13:14:44 although i'm not super happy with it. (tiny) man page is in the binary package :( 2014-06-19 13:16:17 should i change that? subpackage package-doc failed :( 2014-06-19 13:29:52 ScrumpyJack: use `make prefix=... MANDIR="$pkgdir"/usr/share/man/man1 install` and add a -doc subpackage 2014-06-19 13:30:33 the makefile uses /usr/man/man1 which abuild does not detect as documentation 2014-06-19 14:25:37 ah, you want the docs in /usr/share. no problem 2014-06-19 14:33:06 done, thanks again for your help. it should get easier :) 2014-06-19 14:41:00 now send patch to mailing list? 2014-06-19 14:43:30 usually yes 2014-06-19 14:43:40 you can also publish your aports tree with changes you did somewhere 2014-06-19 14:43:44 so we pull it from there 2014-06-19 14:44:19 or if its a single commit you can also: git format-patch -1 --stdout | sprunge 2014-06-19 14:44:30 and paste the url here 2014-06-19 14:44:57 i here will then do: curl $the_url | git am 2014-06-19 14:44:58 yeah, git-email wants perl :( 200+ megs 2014-06-19 14:47:20 ncopa: http://sprunge.us/MgVF 2014-06-19 14:48:19 that is a change of the last commit only 2014-06-19 14:48:25 this is a new package right? 2014-06-19 14:48:25 oops 2014-06-19 14:48:29 yup 2014-06-19 14:48:37 hang on 2014-06-19 14:48:44 you could git rebase -i amd merge the commits 2014-06-19 14:48:47 and* 2014-06-19 14:49:18 i think git rebase -i might depend on perl too... 2014-06-19 14:51:13 http://sprunge.us/QGGf 2014-06-19 14:51:17 is that better? 2014-06-19 15:02:43 looks better. gotta go will look at it tm 2014-06-19 15:05:35 cool. many thanks. 2014-06-20 06:10:52 morning 2014-06-20 06:11:42 trailing whiespace in last patch 2014-06-20 06:12:11 ScrumpyJack: morning, i think we want the 2 commits combined into one 2014-06-20 06:12:14 since its the first 2014-06-20 06:12:22 it makes the history look nicer 2014-06-20 06:15:02 hm 2014-06-20 06:15:14 sc seems to be abandoned upstream... :-/ 2014-06-20 06:15:39 make: bison: Command not found 2014-06-20 06:19:11 ScrumpyJack: do: 'git reset HEAD^' then remove the trailing whitespace in the mkdir line and add bison to makedepends 2014-06-20 06:19:25 finally git add APKBUILD && git commit --amend 2014-06-20 06:19:36 the --amend will make the change go into previous commit 2014-06-20 06:19:56 and change the commit message to something like: testing/sc: new aport 2014-06-20 06:20:00 2014-06-20 06:20:21 2014-06-20 07:48:06 morning 2014-06-20 07:48:53 ah yes, bison is needed 2014-06-20 10:21:06 ncopa: is this better? http://sprunge.us/cWVh 2014-06-20 11:03:51 yes, much better 2014-06-20 11:05:18 um git commit message 2014-06-20 11:05:27 um? 2014-06-20 11:05:51 the 2 first lines shoudl be removed 2014-06-20 11:05:57 ok 2014-06-20 11:06:02 sc (Spreadsheet Calculator) aport for alpine linux 2014-06-20 11:06:07 and the empty line under 2014-06-20 11:06:34 i can fix that here 2014-06-20 11:06:35 is it alpinelinux or alpine linux? 2014-06-20 11:07:21 http://sprunge.us/XbKG 2014-06-20 11:07:25 thanks. so just one line in a commit, with an empty line after it? 2014-06-20 11:07:34 yes 2014-06-20 11:07:53 first line in commit message is considered 'subject' in an email 2014-06-20 11:07:56 then empty line 2014-06-20 11:08:00 and the rest is the body 2014-06-20 11:08:17 i think i just push http://sprunge.us/XbKG 2014-06-20 11:08:51 how exciting! my first alpine package :) 2014-06-20 11:09:05 yay! 2014-06-20 11:09:09 congrats! 2014-06-20 11:09:11 and thanks! 2014-06-20 11:10:31 just making a few more. can I post you the first few patches privately so you can take a look before I commit? (well, i don't have commit, but you know what i mean) 2014-06-20 11:10:46 sure 2014-06-20 11:10:50 thanks 2014-06-20 11:27:42 barthalion: what was the proper fix for bash? 2014-06-20 11:27:46 i think our current is broke 2014-06-20 11:28:02 n. 2014-06-20 11:28:11 ncopa: https://projects.archlinux.org/svntogit/packages.git/tree/trunk/privmode-setuid-fail.patch?h=packages/bash 2014-06-20 11:28:21 this one is made by bash maintainer 2014-06-20 11:28:30 ok 2014-06-20 11:41:00 rebooting brb 2014-06-20 12:06:39 apologies 2014-06-20 13:34:12 ScrumpyJack: i just set pkgrel=0 and pushed it 2014-06-20 13:34:13 thanks 2014-06-20 13:53:39 ncopa: what is the cause of this: denied resource overstep by requesting 46084096 for RLIMIT_MEMLOCK against limit 6553? 2014-06-20 13:53:51 or possible the solution 2014-06-20 13:57:54 ulimit? 2014-06-20 13:57:56 dunno 2014-06-20 13:58:43 checkout /etc/security/limits.conf 2014-06-20 14:00:35 does bb ash respect /etc/security/limits.conf? 2014-06-20 14:01:15 ncopa: thanks 2014-06-20 14:05:10 clandmeter: you were faster than me :) 2014-06-20 14:05:19 of course 2014-06-20 14:05:40 I wonder if that doc subpkg funct makes sence 2014-06-20 16:47:33 there is also http://www.gnu.org/software/recutils/ , still active. 2014-06-20 16:48:01 But nice to have 'sc' 2014-06-20 17:35:25 ncopa: what you think about doing relative not absolute symlinks? so sh in /bin points to busybox not /bin/busybox? 2014-06-20 19:28:14 http://lists.alpinelinux.org/alpine-devel/3827.html 2014-06-20 19:28:19 oh dear 2014-06-20 19:28:40 slashbeast: you should take it up with busybox developers 2014-06-20 19:33:46 kaniini: yeah, but if feature is worth it it could be done via postprocessing. 2014-06-20 19:39:48 it could, but we're not going to do that 2014-06-20 19:41:18 what is the rationale for it 2014-06-20 19:41:36 relative symlinks incur additional overhead for resolution 2014-06-20 19:41:50 and if it's saving 4 bytes per symlink in ramfs, it's not worth it 2014-06-20 19:59:37 usage as prefix-like system, chroot without chroot, with path exported to alpine, using its dynamic loader and LD_LIBRARY_PATH set for lib64 of alpine 2014-06-20 19:59:45 for crosscompilaaption prupose 2014-06-20 19:59:59 but it does resolve symlinks to *my* busybox, not the alpine one 2014-06-20 20:09:08 if you want those things you should use gentoo 2014-06-20 20:31:35 gentoo due to nature of portage is not that useful, in alpine I can just download packages and use them, without bootstraping the base system 2014-06-20 20:31:44 but yeah, whatever, I can hack it on my end. 2014-06-21 21:32:44 this: http://wiki.alpinelinux.org/wiki/Uvesafb doesn't quite work for me. when i boot with this, i get that v86d is not found on the initrd. despite i did run mkinitfs 2014-06-21 21:33:22 so i added /sbin/v86d instead in features.d to an existing and used feature as a quick hack. how would this be done cleanly? 2014-06-22 16:16:03 the alsa devices in /dev/snd are owned by group root and not group audio, how can this be permanently fixed? 2014-06-22 23:51:18 aaah, setting SUBSYSTEM=sound;.*root:audio 660 in /etc/mdev.conf permanently enables audio for users in the audiogroup. neat 2014-06-23 06:12:44 morning 2014-06-23 06:12:58 slashbeast: i'm sceptic to relative symlink to busybox 2014-06-23 06:13:38 WwQCn4HqahlD: do you mind file a bug for the /dev/snd issue? please include your findings how to fix it too 2014-06-23 07:15:25 ncopa: filed bug #3072 with patch for man page compression in abuild 2014-06-23 07:18:14 uggedal: will mdocml-man work with it? 2014-06-23 07:19:39 hmm, will test 2014-06-23 07:21:05 ncopa: the mdocml man wrapper uses zcat/bzcat if the man page is compressed 2014-06-23 07:22:16 i htink it works yes 2014-06-23 07:22:24 brb 2014-06-23 07:24:18 tested it and it works 2014-06-23 07:27:46 ~ du -sh man* 2014-06-23 07:27:46 20.2M man-bz2 2014-06-23 07:27:46 20.4M man-gz 2014-06-23 07:27:47 29.7M man-uncompressed 2014-06-23 07:48:17 applied 2014-06-23 07:48:18 thanks! 2014-06-23 07:51:32 Morning ncopa 2014-06-23 07:51:45 can you remind how you'd like the commit message to look? 2014-06-23 07:52:27 first line like: 2014-06-23 07:52:42 main/package: short summary 2014-06-23 07:52:49 then a blank line 2014-06-23 07:53:13 then an optionally longer description why the change was needed or why it was done they way it was 2014-06-23 07:53:49 you have lots of examples: http://git.alpinelinux.org/cgit/aports/log/ 2014-06-23 07:55:28 you can refer to a bug id in commit message like this: ref # 2014-06-23 07:55:41 and you can resolve bugs with: fixes # 2014-06-23 07:55:54 thanks 2014-06-23 07:57:00 http://tbaggery.com/2008/04/19/a-note-about-git-commit-messages.html 2014-06-23 07:57:21 also wrap lines at 72 columns 2014-06-23 08:52:42 ncopa will do, been told that on another chan some found the libffi6/grsec problem and linked to my report on debian, altought he patch is unrelated in there needs some more investigations. 2014-06-23 08:54:06 ok 2014-06-23 08:54:12 thanks for following it up 2014-06-23 08:54:56 also i earlier had this problem, any answer to this 2014-06-23 08:54:58 23:32:14 WwQCn4HqahlD | this: http://wiki.alpinelinux.org/wiki/Uvesafb doesn't quite work for me. when i boot with this, i get that v86d is not found on the initrd. despite i did run mkinitfs 2014-06-23 08:55:00 23:32:52 WwQCn4HqahlD | so i added /sbin/v86d instead in features.d to an existing and used feature as a quick hack. how would this be done cleanly? 2014-06-23 08:55:27 i think thats the correct thing 2014-06-23 08:55:45 we moved /etc/mkinitfs/file.d/* to /etc/mkinitfs/features.d/*.files 2014-06-23 08:55:53 i'll update the wiki 2014-06-23 08:56:40 ooooh. man. i added /sbin/v86d to i think features.d/base-file.files or similar (cant quite remember the name) 2014-06-23 08:56:53 shouldnt matter that much 2014-06-23 08:57:04 ok. 2014-06-23 08:57:06 thx! 2014-06-23 08:59:36 ncopa: i think the wiki also says mkinitrd instead of mkinitfs 2014-06-23 08:59:54 indeed 2014-06-23 09:01:09 ncopa thanks for your work! 2014-06-23 09:01:27 thanks :) 2014-06-23 09:09:41 ncopa: https://bugs.alpinelinux.org/issues/3073 2014-06-23 09:12:29 good 2014-06-23 09:12:30 thanks! 2014-06-23 09:14:07 we already have some rules for snd 2014-06-23 09:14:09 # alsa sound devices and audio stuff 2014-06-23 09:14:09 pcm.* root:audio 0660 =snd/ 2014-06-23 09:14:09 control.* root:audio 0660 =snd/ 2014-06-23 09:14:09 midi.* root:audio 0660 =snd/ 2014-06-23 09:14:09 seq root:audio 0660 =snd/ 2014-06-23 09:14:10 timer root:audio 0660 =snd/ 2014-06-23 09:14:34 oh, those are for moving the devices to snd/ 2014-06-23 09:14:40 but if kernel already put them there... 2014-06-23 09:17:30 ncopa: morning 2014-06-23 09:17:33 got a min? 2014-06-23 09:17:53 sure 2014-06-23 09:18:14 :) 2014-06-23 09:18:32 im still having timezone issues 2014-06-23 09:18:38 ok? 2014-06-23 09:18:45 https://github.com/ZoneMinder/ZoneMinder/blob/master/src/zm_event.cpp#L76 2014-06-23 09:19:12 this app uses a mix of php perl and c++ 2014-06-23 09:19:25 seems php and perl are using my localtime+timezone 2014-06-23 09:19:47 but the c++ part is using utc 2014-06-23 09:20:31 that is kinda weird 2014-06-23 09:20:47 how do you set localtime? 2014-06-23 09:20:57 with /etc/localrime? 2014-06-23 09:21:01 localtime* 2014-06-23 09:21:08 i use your setup-timezone script 2014-06-23 09:21:21 so it should be /etc/localtime 2014-06-23 09:21:32 does your env set TZ? 2014-06-23 09:21:47 /etc/localtime does not exist 2014-06-23 09:21:54 ok 2014-06-23 09:21:56 thats it then 2014-06-23 09:22:07 so the script is broken? 2014-06-23 09:22:19 i think its fixed now 2014-06-23 09:22:31 do you have /etc/zoneinfo/localtime ? 2014-06-23 09:23:05 lrwxrwxrwx 1 root root 30 Jun 20 10:32 /etc/zoneinfo/localtime -> /etc/zoneinfo/Europe/Amsterdam 2014-06-23 09:23:10 apk upgrade should fix it 2014-06-23 09:23:18 i did 2014-06-23 09:23:23 i suppose this is a box with tmpfs? 2014-06-23 09:23:37 its lxc container 2014-06-23 09:23:38 'diskless' 2014-06-23 09:23:42 hm 2014-06-23 09:23:43 ok 2014-06-23 09:23:44 on 2.7 2014-06-23 09:23:47 err 2014-06-23 09:23:51 not sure its 2.7 2014-06-23 09:23:59 verify that 2014-06-23 09:24:11 v2.7 and v3.0 set timezone very different 2014-06-23 09:24:13 yes host is 2.7 2014-06-23 09:24:17 container 3.0 2014-06-23 09:24:20 ok 2014-06-23 09:24:30 mv /etc/zoneinfo/localtime /etc/localtime 2014-06-23 09:24:32 should fix it 2014-06-23 09:24:47 i though apk upgrade should fix it 2014-06-23 09:25:14 if i do that, date uses UTC again 2014-06-23 09:25:37 date was already ussing the correct timezone. 2014-06-23 09:26:53 is the time on the host correct? 2014-06-23 09:27:10 check also timezone 2014-06-23 09:27:13 yes 2014-06-23 09:27:31 host has correct time and timezone 2014-06-23 09:27:34 date -R 2014-06-23 09:28:10 Mon, 23 Jun 2014 11:27:26 +0200 2014-06-23 09:28:32 Mon, 23 Jun 2014 11:28:24 +0200 2014-06-23 09:28:48 and in container? 2014-06-23 09:29:05 a1:~# date -R 2014-06-23 09:29:05 Mon, 23 Jun 2014 09:28:40 +0000 2014-06-23 09:29:06 Mon, 23 Jun 2014 09:28:57 +0000 2014-06-23 09:29:18 that is correct 2014-06-23 09:29:29 i mean the time is correct 2014-06-23 09:29:42 yes but without timezone 2014-06-23 09:29:56 since i moved the file you told me, it displays in correct now 2014-06-23 09:30:18 remeember we had an issue with syslog? 2014-06-23 09:30:24 i guess it also logs incorrectly for you 2014-06-23 09:30:34 a1:~# date -R 2014-06-23 09:30:34 Mon, 23 Jun 2014 11:30:23 +0200 2014-06-23 09:30:46 i get that after setup-timezone 2014-06-23 09:31:12 yes me 2 2014-06-23 09:31:25 then it is ok? 2014-06-23 09:31:29 but still c/c++ is using wrong localtime (i guess locatime should have timezone mod) 2014-06-23 09:32:00 maybe syslog needs to be restarted? 2014-06-23 09:32:12 after moving /etc/zoneinfo/localtime 2014-06-23 09:33:00 Jun 23 09:32:27 zoneminder syslog.info syslogd exiting 2014-06-23 09:33:00 Jun 23 11:32:27 zoneminder syslog.info syslogd started: BusyBox v1.22.1 2014-06-23 09:33:00 Jun 23 09:32:40 zoneminder user.notice root: test 2014-06-23 09:33:29 is it logging correctly for you in syslog? 2014-06-23 09:33:31 how do you generate that log? 2014-06-23 09:33:39 with logger 2014-06-23 09:33:54 Jun 23 11:31:49 a1 syslog.info syslogd started: BusyBox v1.22.1 2014-06-23 09:33:54 Jun 23 09:33:45 a1 user.notice root: test 2014-06-23 09:34:31 i think all c/c++ based bins do not use timezone correctly 2014-06-23 09:36:53 https://gist.github.com/ncopa/68c6fd8471a5a2de706d 2014-06-23 09:38:48 i think it could be a bug in 'logger' 2014-06-23 09:39:09 https://gist.github.com/ncopa/a9eb06c199c3f8831692 2014-06-23 09:39:19 could be, but then its also a bug in zoneminder 2014-06-23 09:41:35 https://github.com/ZoneMinder/ZoneMinder/blob/master/src/zm_event.cpp#L76 2014-06-23 09:41:45 it should detect the localtime 2014-06-23 09:42:12 and then it calls mysql: from_unixtime( %ld ) 2014-06-23 09:42:12 it should 2014-06-23 09:42:15 but it doesnt 2014-06-23 09:42:22 it inserts utc in mysql 2014-06-23 09:43:54 storing the utc timestamp in db sounds like a good idea 2014-06-23 09:44:22 but then it shoudl be configurable which is the preferred presentation timezone 2014-06-23 09:44:23 true 2014-06-23 09:44:36 im seeing the code 2014-06-23 09:44:41 it aint pretty 2014-06-23 09:44:44 :) 2014-06-23 09:45:03 it creates a dir sturct based on local time 2014-06-23 09:45:10 but the db has utc 2014-06-23 09:45:17 is timezone of timestamp stored in db? 2014-06-23 09:45:18 so operations on it will fail 2014-06-23 09:45:50 no, afaik its all taking system timezone 2014-06-23 09:45:52 maybe you need restart mysql after the /etc/localtime change? 2014-06-23 09:46:29 I can 2014-06-23 09:46:45 but that line i gave you doesnt use mysql date functions. 2014-06-23 09:47:11 ugh 2014-06-23 09:47:18 looks like mysql is stupid 2014-06-23 09:47:29 oh maybe it does. 2014-06-23 09:56:57 ncopa: you are right 2014-06-23 09:57:06 when i restarted the container, time is ok. 2014-06-23 10:00:20 ok good :) 2014-06-23 11:27:18 this is what i have been hacking at: pampers-0.1-beta: https://eu.mirrors.as250.net/misc/qv.txz alpinelinux/qemu-kvm pdf & mplayer compartment, framebuffer no x11, no systemd, no glibc. 2014-06-23 11:28:58 jfbview, mupdf-nox11 and an updated openjpeg are available from my aports at github/stef/aports 2014-06-23 11:51:23 WwQCn4HqahlD: you replaced mupdf with your mupdf-nox11 2014-06-23 11:55:02 jfbview is built using git master, you should use a versioned tarball instead 2014-06-23 11:55:10 probably store it somewhere 2014-06-23 11:55:32 zfs is magic, so I leave that to ncopa :p 2014-06-23 11:59:03 is fortify.patch missing? I can't check because I'm behind 56k modem and I'm gonna die in a moment due to slow transfers 2014-06-23 11:59:27 I guess it is 2014-06-23 11:59:42 makedepends='git mupdf-nox11-dev imlib2-dev jpeg-dev openjpeg-dev openssl-dev jbig2dec-dev@testing' 2014-06-23 11:59:47 you can dump @testing 2014-06-23 12:00:50 additionally you bumped openjpeg to 2.1.0, which has incompatible API to previous release 2014-06-23 12:01:19 so it will indeed work for mupdf, but will also make apk unhappy thanks to changed soname 2014-06-23 12:04:53 in Arch we provide openjpeg 2.x as openjpeg2, as literally nothing uses it except mupdf 2014-06-23 12:19:49 ooh shit, yes, fortify patch adding. 2014-06-23 12:20:10 oooh, kool 2014-06-23 12:20:24 will adress these things. thx! 2014-06-23 12:20:55 btw fortify is just cflags with stackprotect et al 2014-06-23 12:23:48 yw 2014-06-23 12:59:02 barthalion: when i search for mupdf on packages.a.o i don't get any result, was mupdf packaged for alpine linux? 2014-06-23 13:16:39 ncopa: small typo patch for abuild: http://sprunge.us/NLOV 2014-06-23 13:23:05 barthalion: also jfbview uses openjpeg2 so i guess i'll rename it an package it like arch 2014-06-23 13:48:06 uggedal: thanks! 2014-06-23 13:48:35 i suppose we have enough for a new .dot release 2014-06-23 13:49:16 WwQCn4HqahlD: search on packages.a.o doesn't use testing 2014-06-23 13:49:23 ooooh 2014-06-23 13:50:47 you basically replaced testing/mupdf/APKBUILD content with mupdf-x11 of yours 2014-06-23 13:51:04 lemme check 2014-06-23 13:53:20 ooops 2014-06-23 13:55:29 ok, i'll clean this up and revert to the original mupdf and stuff. i have no idea how i missed this 2014-06-23 13:58:35 thanks for the review! 2014-06-23 14:08:15 barthalion: i see, that mupdf in testing was only 1.0 so my update to 1.4 makes sense actually no? 2014-06-23 14:13:26 WwQCn4HqahlD: as long as it's a update, not major overhaul with different set of features and name 2014-06-23 14:15:15 ncopa, isn't #3060 fixed ? 2014-06-23 14:15:46 WwQCn4HqahlD: and iirc, I've recently pushed 1.5 to Arch… 2014-06-23 14:15:56 ooh i'll update that then. 2014-06-23 14:16:02 fcolista: i normally keep the issue open til its fixed in the target branch 2014-06-23 14:16:06 eg v3.0 git branch 2014-06-23 14:16:15 oh, ok 2014-06-23 14:16:16 fine 2014-06-23 14:16:24 i tought that something was still wrong 2014-06-23 14:16:37 otheriwsei close it, and then i look in the v3.0.1 target issue list 2014-06-23 14:16:42 and think "everything is ok" 2014-06-23 14:16:45 and ship a release 2014-06-23 14:17:17 ok 2014-06-23 14:18:07 i really miss the multiple target branches feature in redmine 2014-06-23 14:19:22 yeah, that would have been useful 2014-06-23 14:19:24 does jira have it? 2014-06-23 14:19:42 we could probably switch then, atlassian gives free licences to open source projects 2014-06-23 14:20:07 https://answers.atlassian.com/questions/10110/what-is-the-best-way-to-handle-fix-targets-for-multiple-branches 2014-06-23 14:20:26 bah 2014-06-23 14:20:28 that's old 2014-06-23 14:21:04 everything starting with a j sucks even more :p 2014-06-23 14:21:09 lol 2014-06-23 14:22:17 on the bright side, we wouldn't need to pray before ruby upgrade :p 2014-06-23 14:22:39 +1 2014-06-23 14:24:37 how about flyspray? 2014-06-23 14:25:05 i'd love avoid the ruby upgrade prayers... 2014-06-23 14:25:32 Flyspray 0.9.9.7 - 28 May 2012 2014-06-23 14:25:37 looks unmaintained 2014-06-23 14:25:46 yeah, we want to switch from it in Arch 2014-06-23 14:25:59 good to know 2014-06-23 14:26:03 any good candidates? 2014-06-23 14:26:08 possibly to bugzilla if someone invent the script to ease migration 2014-06-23 15:00:34 do we care about this? https://bitbucket.org/GregorR/musl-pkgsrc-patches 2014-06-23 15:01:56 maybe idividual patches 2014-06-23 15:02:42 but i think we have most of them working already 2014-06-23 15:02:47 https://bitbucket.org/GregorR/musl-pkgsrc-patches/src 2014-06-23 15:03:13 I'm trying to compile wmi-client, but it complains about setnetgrent that is not available 2014-06-23 15:03:32 this should add support for setnetgrent 2014-06-23 15:03:41 this = this set of patches 2014-06-23 15:04:16 i gotta go see u 2014-06-23 15:04:25 bye 2014-06-23 15:05:56 bah...grep -r setnetgrent return anything 2014-06-23 18:49:15 huh, i try rework the jfbview dependencies but now i get a strange "not committing changes due to missing repository tags" when doing abuild -r - what is that about? 2014-06-23 19:02:52 ok i got that, some package in testing uses @testing as their deps. 2014-06-23 19:04:13 WwQCn4HqahlD: try to manually add the dependencies with apk add -v 2014-06-23 19:05:13 WwQCn4HqahlD: nevermind, skipped your last msg :) 2014-06-23 19:06:08 thx anyway! 2014-06-23 19:54:12 hmm, anyone using abuild with source from local disk (out of tree) 2014-06-23 19:54:42 i've tried relative paths, but abuild checksum and abuild fetch does not use the same base directory 2014-06-23 19:55:10 i could add file:/// support to abuild, but the busybox wget fallback does not support that protocol 2014-06-23 19:56:09 I tried it one day, but didn't really get anywhere 2014-06-23 19:56:58 this is for a private repo, I could spin up a httpd on localhost to serve the source, but I'd rather not... 2014-06-23 19:57:00 checking if given element of source array is a directory and copying it to $srcdir would do I guess 2014-06-23 19:58:57 I would be ok with relative or absolute paths 2014-06-23 19:59:11 actually abuild checksum and fetch uses the same base, but unpack does not 2014-06-23 20:01:16 we should just make alpine-sdk pull in curl 2014-06-23 20:04:25 well, wouldent that be bad for bootstrap? 2014-06-23 20:04:54 conditonal file:// support based on curl presence would work for me 2014-06-23 20:36:30 heh, found a hack for now: source="saveas-file://$HOME/src/mediaqueri.es/dist/$pkgname-$pkgver.tar.gz" 2014-06-23 23:05:38 so i updated my aports repo with clean patches for each package, i tried to follow up on all feedback earlier today, any more feedback is welcome, cherrypick packages: https://github.com/stef/aports/commits/master 2014-06-24 06:06:00 uggedal: i think you can specify a distfiles mirror where it will look for all files 2014-06-24 06:06:43 set DISTFILES_MIRROR=/local/path/to/dir in abuild.conf 2014-06-24 06:06:55 and it will look there before it tries to wget it 2014-06-24 06:08:21 i think you also can create a local fetch() function in your APKBUILD 2014-06-24 06:09:39 fetch() { mkdir "$srcdir"; for _s in $source; do ls -s $s "$srcdir"/; } 2014-06-24 08:52:36 ncopa: would it be possible to patch abuild with file:// uri support? 2014-06-24 08:53:41 ncopa: will you fix setup-timezone? or should it have a bug assigned? 2014-06-24 08:56:28 I'm trying to build some private packages where I don't want to serve the sources up over http 2014-06-24 08:57:00 I'm currently using this hack: source="saveas-file://$HOME/src/mediaqueri.es/dist/$pkgname-$pkgver.tar.gz/$pkgname-$pkgver.tar.gz"... 2014-06-24 08:57:19 clandmeter: whats wrong with it? please file a bug 2014-06-24 08:57:36 uggedal: yes it is possible... 2014-06-24 08:57:38 ncopa: what happend yesterday? 2014-06-24 08:57:43 the issue i had. 2014-06-24 08:58:00 uggedal: i wonder if we want: file:// or simply /path/to/file 2014-06-24 08:59:46 http://sprunge.us/FhjZ 2014-06-24 09:00:32 that patch should let you specify a local absolute path 2014-06-24 09:02:45 uggedal: this should add support for both file:///path/to/file and /path/to/file: http://sprunge.us/DfUU 2014-06-24 10:31:50 ncopa: thanks. the first patch does not work since default_unpack extracts from $SRCDEST 2014-06-24 10:32:36 basically non-uri sources can not be archives 2014-06-25 06:05:07 morning all 2014-06-25 06:07:24 morning 2014-06-25 06:54:57 I'm thinking of packaging uglifyjs and cleancss and their node module dependencies 2014-06-25 06:55:08 what do we want as prefix for node packages? 2014-06-25 06:55:18 node-async, nodejs-async, js-async? 2014-06-25 06:55:24 debian uses node- 2014-06-25 06:56:07 and fedora prefixes with? 2014-06-25 06:56:33 nodejs 2014-06-25 06:56:35 nodejs-* 2014-06-25 06:56:54 uggedal: i dont have any strong opinion if you do node-* or nodejs-* 2014-06-25 06:57:39 but i think i'd prefer either node-* or nodejs-* 2014-06-25 07:09:59 ncopa: alpine 3.0 doesnt boot on vmware esxi. 2014-06-25 07:10:06 i had to use the vanilla iso 2014-06-25 07:10:19 but that iso is also weird 2014-06-25 07:10:38 setup-alpine will install grsec kernel when booting vanilla 2014-06-25 07:11:22 and timezone is not correct on 3.0 2014-06-25 07:11:42 apk upgrade should fix 2014-06-25 07:11:50 and fix should be in 3.0.1 2014-06-25 07:12:07 yes but its broken for 3.0 2014-06-25 07:12:19 so anybody useing 3.0 has wrong timezone 2014-06-25 07:12:32 fabled: i think we want push the musl update for /etc/localtime soonish 2014-06-25 07:13:02 ncopa, if dalias tags new musl today, i'd hope to push that 2014-06-25 07:13:48 ncopa: i used setup-disk with -k vanilla, and that breaks extlinux conf. 2014-06-25 07:15:44 clandmeter: i just pushed a fix for update-extlinux and vanilla kernel 2014-06-25 07:16:17 ncopa: setup-timezone should keep /etc/zoneinfo/localtime to be backwards compat for the time beeing. 2014-06-25 07:21:25 Morning 2014-06-25 07:33:57 probably stupid question, but how do i edit the boot parameters of the alpine iso? 2014-06-25 07:34:08 at boot? 2014-06-25 07:34:11 2014-06-25 07:34:26 grsec 2014-06-25 07:35:24 i want to boot without quiet 2014-06-25 07:35:35 2014-06-25 07:35:40 grsec noquiet 2014-06-25 07:36:41 or that doesnt work, or it crashes really early 2014-06-25 07:37:00 want there another debug option? 2014-06-25 07:37:04 wasnt 2014-06-25 07:37:43 grsec noquiet verbose pax_nouderef 2014-06-25 07:37:57 i think you need pax_nouderef on vmware 2014-06-25 07:39:45 doenst do anything :| 2014-06-25 08:05:15 ncopa: any idea about this one? http://sprunge.us/FhfJ 2014-06-25 08:07:01 ncopa, seems no new musl today. do you want 3.0.1 out today? if yes, we need to just cherry-pick the commit. 2014-06-25 08:07:16 though musl tagging is imminent. probably tomorrow. 2014-06-25 08:18:06 ncopa: when i boot with 1 cpu it works, adding more will result in the previous oops 2014-06-25 08:18:36 fabled: i'm just afraid of new kernel releases before we reach ship 3.0.1 2014-06-25 08:18:50 i think we should probably cherry-pick the commit 2014-06-25 08:28:51 just wondering if it would be useful to have another folder in aports - "aports/extras" ? 2014-06-25 08:29:27 this would be for packages is nice to have but not bother to put in main 2014-06-25 08:29:29 eg. polarssl 2014-06-25 08:33:58 how would you decide what should go in main vs extras? 2014-06-25 08:34:24 i have been thinking of it though 2014-06-25 08:34:37 becuase the maintenance burden is fairly big already 2014-06-25 08:35:35 it would have weight between main and testing 2014-06-25 08:35:52 packages that maybe be maintained but less prio 2014-06-25 08:36:30 it can also serve as place for ui pkgs 2014-06-25 08:37:07 I mean gtk, kde etc apps 2014-06-25 08:37:09 you could group the aports by category (eg ui apps and cli apps) 2014-06-25 08:37:45 you could group by maintainer (eg 'core' vs 'extras' and i would only care about 'core') 2014-06-25 08:38:09 the bsd ports and gentoo does by category 2014-06-25 08:38:25 I perceive AL more cli oriented / specialized than x/gui 2014-06-25 08:38:49 i'm not sure i like that because its not always clear what category you should put the app in 2014-06-25 08:38:53 I think this should be its selling point 2014-06-25 08:39:31 often aports belongs to multiple categories 2014-06-25 08:39:36 for example, wpa_supplicant 2014-06-25 08:39:54 same aport builds the deamon, a cli client 2014-06-25 08:39:58 and a qt app 2014-06-25 08:40:00 its cli based and has a gui too 2014-06-25 08:40:05 correct 2014-06-25 08:40:06 so can be split 2014-06-25 08:40:25 i dont want multiple APKBUILDs for wpa_supplicant 2014-06-25 08:40:46 in some cases i do want split 2014-06-25 08:40:52 hmm 2014-06-25 08:40:53 like poppler and avahi 2014-06-25 08:41:03 i have asked avahi devs to split it upstream 2014-06-25 08:41:59 my point is 2014-06-25 08:42:22 its not always easy to know in what dir the aports belongs 2014-06-25 08:42:39 which is why i have been doing one big 'main' 2014-06-25 08:43:03 it would be so nice if devs build libs is such a way that core lib (cli) and gui stubs can be build separately 2014-06-25 08:43:06 we split by state more than by category 2014-06-25 08:43:23 yes, many times it woudl be nice 2014-06-25 08:43:34 but very often its not like that 2014-06-25 08:43:46 very often there is a ./configure option for --enable-gui 2014-06-25 08:43:50 and thats it 2014-06-25 08:44:53 may crazy option like ./configure --build-gui-only 2014-06-25 08:51:21 that completely depends on what upstream does 2014-06-25 08:51:36 cannot expect that all upstream packagers do what we prefer 2014-06-25 08:52:44 ncopa: new kernels are under review and I guess they'll be out today 2014-06-25 08:54:05 if we are re-evaluating tracker app, then there is also roundup-tracker.org 2014-06-25 08:54:40 don't know if redmine data can be imported into it 2014-06-25 08:58:49 urgh... :) 2014-06-25 09:00:08 we are too slow... 2014-06-25 11:05:43 re: bug-tracker..in testing there's flyspray available...i've also created a simple wiki page: http://wiki.alpinelinux.org/wiki/Flyspray#Flyspray_configuration 2014-06-25 11:06:21 i've tested it both with pgsql and mysql 2014-06-25 11:06:55 fcolista: afaik archlinux is trying to move away from flyspray 2014-06-25 11:07:02 i dunno the reason 2014-06-25 11:07:10 is not mantained 2014-06-25 11:07:17 *no longer 2014-06-25 11:07:32 ok 2014-06-25 11:07:32 latest version uses deprecated libs 2014-06-25 11:07:42 we cannot use that then 2014-06-25 11:07:57 i think so, but was a nice exercise :) 2014-06-25 11:08:03 :) 2014-06-25 11:08:19 regarding splitting of main, exherbo uses both different repos and categories across them: http://git.exherbo.org/summer/ 2014-06-25 11:08:35 yes 2014-06-25 11:08:41 thats also a possibility 2014-06-25 11:08:45 to split repos 2014-06-25 11:09:01 but that makes it even more confusing for users 2014-06-25 11:09:10 heh, yeah :) 2014-06-25 11:09:15 if i want package 'foo', which repo should I add? 2014-06-25 11:09:46 true 2014-06-25 11:09:50 splitting repos would make the maintenance part easier though 2014-06-25 11:09:58 I like arch's approach 2014-06-25 11:09:59 you maintain this repo and i that 2014-06-25 11:10:18 i kinda like arch's approach too 2014-06-25 11:10:26 and AUR 2014-06-25 11:10:36 yeah, AUR is getting git support 2014-06-25 11:10:37 well, our approach is mainly to prevent unwanted fuckups on whole distro level 2014-06-25 11:10:58 :) 2014-06-25 11:11:18 and it makes sense only due to number of contributors 2014-06-25 11:11:42 how many active devs in terms of packaging do we have now? 6, 7 at best 2014-06-25 11:11:58 thats not much better than alpine 2014-06-25 11:12:07 I mean in Alpine 2014-06-25 11:12:12 ok :) 2014-06-25 11:12:14 yes 2014-06-25 11:13:57 what i'd like to do is delegate responsability to many 2014-06-25 11:14:12 so one could have reponsability of one subsystem 2014-06-25 11:14:18 and another has something else 2014-06-25 11:14:42 but at the same time not make the internal organization too visible to end users 2014-06-25 11:14:47 what about per component repos automerged to the aports repo? 2014-06-25 11:14:55 ↑ 2014-06-25 11:14:58 end users only see the merged repo 2014-06-25 11:15:09 yes, something like that 2014-06-25 11:15:35 it would partially solve issues related to mess in testing, I like it 2014-06-25 11:16:01 testing is actually not so much about dividing reponsability 2014-06-25 11:16:14 its more for basic QA 2014-06-25 11:16:50 prevent total distro level meltdowns 2014-06-25 11:16:51 I mainly mean lots of unmaintained packages there 2014-06-25 11:16:54 yes 2014-06-25 11:16:57 usually outdated and not building 2014-06-25 11:17:25 having component repos would help a bit, as things would be move faster 2014-06-25 11:17:30 or not added at all 2014-06-25 11:17:56 a bit like what linus does with linux 2014-06-25 11:18:02 have sub maintainers 2014-06-25 11:18:46 but at the same time it would remove a bit of the 'rolling release' 2014-06-25 11:20:01 in favour of stability I guess 2014-06-25 11:20:35 maybe basing on gentoo's categories is too much, but things like core, x11, python, ruby could be separated 2014-06-25 11:20:42 and the rest kept the old way 2014-06-25 11:23:30 I recently though about kind of mirror monitoring system 2014-06-25 11:24:02 Arch has it integrated into developer panel, I haven't seen any ready solution 2014-06-25 11:24:38 maybe writing something in flask can be good, as we don't need much more than indicators of mirror status and its last sync 2014-06-25 11:24:50 and maybe notifications if something is behind or down 2014-06-25 11:32:20 what are the main requirements for our nex bug trackings system? 2014-06-25 11:32:25 *next 2014-06-25 11:32:30 http://en.wikipedia.org/wiki/Comparison_of_issue-tracking_systems 2014-06-25 11:32:49 Looks that Jira is one of the best candidates 2014-06-25 11:32:58 but needs java 2014-06-25 11:33:09 and is non-free 2014-06-25 11:33:17 jira is free-as-in-beer for open source 2014-06-25 11:33:20 there's a license ofr free 2014-06-25 11:33:21 we need support for sub-tickets and/or support for multiple targets 2014-06-25 11:33:48 support for closing/resolving tickets by git commits 2014-06-25 11:34:01 what are sub-tickets? The possibility to create a ticket from another ticket? 2014-06-25 11:34:27 you can create ticket X, for fixing X you need fix Y and Z 2014-06-25 11:34:33 ah 2014-06-25 11:34:34 and Y and Z are different tickets 2014-06-25 11:34:47 so ticket dependencies 2014-06-25 11:34:52 ok, fine 2014-06-25 11:34:57 I actually know someone who could host jira for us 2014-06-25 11:35:03 (if we decide) 2014-06-25 11:35:13 i am kinda ok with redmine 2014-06-25 11:35:22 except missing features? :p 2014-06-25 11:35:24 the only things are upgrades are scary 2014-06-25 11:35:31 yeah… 2014-06-25 11:35:33 and missing feature for multiple targets 2014-06-25 11:35:37 did someone touch ruby yet? 2014-06-25 11:36:18 doesn't look like 2014-06-25 11:36:28 its scary remember :) 2014-06-25 11:36:31 nobody dares :) 2014-06-25 11:36:56 hm, new patchlevel for our branch is out 2014-06-25 11:37:32 re ruby packages 2014-06-25 11:37:37 I'll bump it in a moment 2014-06-25 11:37:40 actually is quite difficult to find those requirements explicitly mentioned 2014-06-25 11:37:55 i think it would e nice to test things at package build time 2014-06-25 11:38:00 like perl does 2014-06-25 11:39:37 re maintaining packages 2014-06-25 11:40:07 the fundamental problem is not enough active skilled devs which can take responsability 2014-06-25 11:40:57 and that problem cannot really be solved by dividing/reorganizing the tree structure 2014-06-25 11:45:18 on the other side, it'll make a bit easier to contribute for people who doesn't exactly know how to help 2014-06-25 11:45:23 s/how/where/ 2014-06-25 11:48:12 right 2014-06-25 11:48:22 so we want make it easy to contribute 2014-06-25 11:48:31 we want make it easy and fun to contribute 2014-06-25 11:48:42 well, if the problem is untouched/unmaintained packages (like we saw leading up to 3.0) I think moving loads of packages out of main/testing could help 2014-06-25 11:49:18 if unmaintained was built and available on mirrors lots of packages could be moved there from main/testing 2014-06-25 11:49:34 hm 2014-06-25 11:49:50 basically, that is what 'testing' is in practice 2014-06-25 11:50:04 and when things no longer build we move it to unmaintained 2014-06-25 11:50:09 where it will rot... 2014-06-25 11:50:11 yeah, but then testing could be used for its intended purpose 2014-06-25 11:50:29 true 2014-06-25 11:50:42 so, a package can only stay in testing for X days/weeks/months 2014-06-25 11:51:00 or til it breaks the building 2014-06-25 11:51:05 a treeclean-like script could be run which moved untouched packages in testing to unmaintained 2014-06-25 11:51:06 3 months max I guess 2014-06-25 11:51:25 that sounds pretty good actually 2014-06-25 11:51:45 if no one bothered to test it for 3 months, whether it's packager or reporter, it should be dropped 2014-06-25 11:51:54 agree 2014-06-25 11:51:54 yes 2014-06-25 11:52:16 and maybe we can just purge it... 2014-06-25 11:52:28 instead of moving it to 'trashcan' :) 2014-06-25 11:52:34 (aka unmaintained) 2014-06-25 11:52:38 git has this great feature called history :) 2014-06-25 11:52:43 exactly :) 2014-06-25 11:52:45 if we need to pull back something 2014-06-25 11:53:02 keeping rocket science APKBUILDs could be nice 2014-06-25 11:53:14 the only reason i move things to unmaintained is that it becomes more visible 2014-06-25 11:53:31 unmaintained without purging works for me 2014-06-25 11:54:05 we could purge stuff from unmaintained too on time basis 2014-06-25 11:55:34 anyone has a clue what to do with: http://bugs.alpinelinux.org/issues/2773 2014-06-25 11:55:42 baiscally its an ipv6 feature 2014-06-25 11:55:43 DAD 2014-06-25 11:56:18 debian's ifupdown fixed it by running a wait script on ifup for inet4 2014-06-25 11:56:22 inet6* 2014-06-25 11:56:26 but not inet4 2014-06-25 11:56:37 DAD is inet6 only 2014-06-25 11:56:45 i think they patchs ifupdown.c so they dont need run the script for ipv4 2014-06-25 11:56:56 DAD = Duplpicate address detection 2014-06-25 11:57:03 i wonder if we should patch busybox similar 2014-06-25 11:57:07 it's the time address is configured, but not usable yet 2014-06-25 11:57:28 basically the kernel will probe network with icmpv6 to check that no one else is using the address already 2014-06-25 11:57:54 and meanwhile, ifup has already exited with 'success' but when service wants to bind to that address it fails 2014-06-25 11:57:55 i think the specs say it's more or less mandatory step 2014-06-25 11:59:03 this kind of problem should ideally be solve with kevent based init system (ala systemd) 2014-06-25 12:00:09 so i think we simply port this to busybox? http://anonscm.debian.org/hg/collab-maint/ifupdown/rev/a7e1677e7efc 2014-06-25 12:08:59 algitbot: yay! 2014-06-25 12:09:03 \o/ 2014-06-25 12:09:22 algitbot: hello 2014-06-25 12:09:25 hi 2014-06-25 12:09:26 hello 2014-06-25 12:09:38 w00t! 2014-06-25 12:09:46 wut 2014-06-25 12:09:49 algitbot: c'mon 2014-06-25 12:09:51 \o/ 2014-06-25 12:09:53 ha! 2014-06-25 12:31:21 argh 2014-06-25 12:31:27 i hate ruby on rails 2014-06-25 12:31:52 so we have this security upgrade for ruby-actionpack 4.0.4 to 4.0.5 2014-06-25 12:32:14 i upgrade anything that has 4.0.4 in ruby-*/APKBUILD 2014-06-25 12:32:27 and to verify i have not broken anything i try rails: 2014-06-25 12:32:46 $ PATH=$PATH:/usr/lib/ruby/gems/2.0.0/gems/rails-4.0.5/guides/code/getting_started/bin rails new myapp 2014-06-25 12:32:52 Could not find rake-10.1.0 in any of the sources 2014-06-25 12:32:52 Run `bundle install` to install missing gems. 2014-06-25 12:34:14 oh 2014-06-25 12:34:21 rake is shipped with ruby itself 2014-06-25 12:34:23 great... 2014-06-25 12:36:02 oh 2014-06-25 12:36:06 not only that 2014-06-25 12:36:16 we dont only have ruby-actionpack 2014-06-25 12:36:35 we also have ruby-redmine-actionpack 2014-06-25 12:36:40 that also needs security fix 2014-06-25 12:36:45 argh1 2014-06-25 12:48:32 ncopa: mind if I take over maintainership of nodejs? 2014-06-25 12:48:44 please do, thanks! 2014-06-25 13:09:51 would you mind awfully keeping unmaintained for a bit? i wouldn't mind having a poke about and seeing what I can pick up and refresh 2014-06-25 13:11:38 ScrumpyJack: thats the reason it exists 2014-06-25 13:11:47 alpine 2.5 is still supported? 2014-06-25 13:11:50 i dont mind keeping it around 2014-06-25 13:11:57 Xe: good question 2014-06-25 13:12:01 in theory yes 2014-06-25 13:12:11 but i am not able to push kernel fixes for it 2014-06-25 13:12:27 why not? 2014-06-25 13:12:38 because the kernel is out of support upstream 2014-06-25 13:12:44 we picked a bad kernel for it 2014-06-25 13:12:51 ah 2014-06-25 13:12:53 lol 2014-06-25 13:12:59 and they dont announce in advance which will be LTS.. 2014-06-25 13:13:51 would downgrade to 3.4 hurt much? 2014-06-25 13:14:28 ncopa: what versions of Alpine are currently supported? 2014-06-25 13:22:36 Xe: http://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases 2014-06-25 13:23:59 thanks 2014-06-25 17:55:58 hi@all 2014-06-25 17:56:04 is there something broken with shared-mime-info? 2014-06-25 17:56:07 because: 2014-06-25 17:56:16 (1/1) Installing shared-mime-info (1.2-r2) 2014-06-25 17:56:16 0% [ ] 2014-06-25 17:56:16 wget: short write 2014-06-25 17:56:16 ERROR: shared-mime-info-1.2-r2: BAD signature 2014-06-25 17:56:16 1 errors; 575 MiB in 153 packages 2014-06-25 17:56:33 StarWarsFan: df -h? 2014-06-25 17:57:55 http://sprunge.us/SDJg 2014-06-25 17:59:57 I encountered such errors when I had no space left on the partition 2014-06-25 18:00:05 not much help here 2014-06-25 18:00:14 that seems to be not the case in that case... 2014-06-25 18:14:33 hmm missing pkgs on master mirror 2014-06-25 18:16:33 again? 2014-06-25 18:17:23 i dunno what happend 2014-06-25 18:17:40 zoneminder is missing even if it says uploaded 2014-06-25 18:18:40 algitbot: you little liar 2014-06-25 18:18:44 lol 2014-06-25 18:24:58 algitbot: upload 2014-06-25 18:25:11 its probably not that smart 2014-06-25 18:28:55 if it fails to build, it'll be just deleted, but algitbot still says uploaded. there's still too many failing pkgs in musl build 2014-06-25 18:29:19 git://git.alpinelinux.org/tteras/quagga 2014-06-25 18:29:26 oops 2014-06-25 18:29:28 meant: http://bld1.alpinelinux.org/buildlogs/build-edge-x86/main/zoneminder/zoneminder-1.27.0-r9.log 2014-06-25 18:29:38 per-pkg build logs are at http://bld1.alpinelinux.org/buildlogs/build-edge-x86/ 2014-06-25 18:34:57 fabled: hi 2014-06-25 18:35:03 thx for that link 2014-06-25 18:37:03 x86 still did not build (x86_64 built) 2014-06-25 18:37:05 zm_utils.cpp: In function 'void ssedetect()': 2014-06-25 18:37:05 zm_utils.cpp:208:3: error: PIC register clobbered by '%ebx' in 'asm' 2014-06-25 18:47:38 has that to do with the bad-sig-thing above? 2014-06-25 18:54:31 StarWarsFan, no, BAD sig like that is usually due to mirror being out of sync 2014-06-25 18:54:44 try another mirror 2014-06-25 18:54:48 ok 2014-06-25 18:55:36 and if i got you here: did you get any idea on the nfs-issue? some days ago ncopa count on you... 2014-06-25 19:12:30 StarWarsFan, sorry, i've been way too busy. things settle down after a week or so. 2014-06-25 19:30:56 fabled: ok, thx for the info 2014-06-26 06:00:13 err, moved from testing... 2014-06-26 06:19:47 ncopa, re #3090. setup-dnscrypt is a script that is not shipped with the package, but the issue ask to add sed and curl among the dependencies. Should i add it? Or maybe is better that the script ask for installing it with a check, before it run? 2014-06-26 06:35:17 Morning 2014-06-26 07:13:58 calibre disuade linux distro packaging http://calibre-ebook.com/download_linux 2014-06-26 07:14:39 should an aport be maintained in that case? 2014-06-26 07:15:01 (calibre is currently in unmaintained) 2014-06-26 07:18:28 fcolista: why is GNU sed needed? 2014-06-26 07:18:58 ncopa, is needed by setup-dnscrypt, but i think that busybox sed is enough 2014-06-26 07:19:21 bb sed should be enough, if its not, i'd suggest to fix the sed line so it works with bb sed 2014-06-26 07:19:23 re curl 2014-06-26 07:19:29 we could use wget -O - ? 2014-06-26 07:19:34 is https 2014-06-26 07:19:37 ok 2014-06-26 07:19:51 so we can chose between curl or wget 2014-06-26 07:19:53 :) 2014-06-26 07:20:01 is this script a separate aport 2014-06-26 07:20:05 but, again, problems are related to the script 2014-06-26 07:20:10 os is script shipped with other package? 2014-06-26 07:20:32 is in the same package. but is homebrew 2014-06-26 07:20:42 which package is it? 2014-06-26 07:20:46 lstuart used to do it 2014-06-26 07:20:52 dnsproxy-crypt 2014-06-26 07:20:58 dnscrypt-proxy 2014-06-26 07:21:34 i suppose we could ship the setup script in separate package? 2014-06-26 07:21:38 i seek another source for the csv script download.. 2014-06-26 07:21:46 good point 2014-06-26 07:21:48 which depends on curl 2014-06-26 07:21:55 subpkg would be ok? 2014-06-26 07:22:12 i think so 2014-06-26 07:22:20 then we dont need depend on curl for the main package 2014-06-26 07:22:26 Fine. What should be the name? 2014-06-26 07:22:31 $pkgname-setup? 2014-06-26 07:22:56 umh 2014-06-26 07:23:04 ah 2014-06-26 07:23:05 ok 2014-06-26 07:23:27 i'm just wondering that sbin/setup-* are script shipped with alpine 2014-06-26 07:23:31 so 2014-06-26 07:23:51 setup-dnscrypt in /sbin is not a nice place/name imho 2014-06-26 07:24:10 also the quality of the script is not the same. 2014-06-26 07:24:23 would be better to port it in lua.. 2014-06-26 08:09:48 i havent studied the script but i got the feeling it needs a cleanup yes 2014-06-26 08:09:59 or rewrite... 2014-06-26 08:11:16 fabled: so we are more or less ready for 3.0.1? 2014-06-26 08:11:45 ncopa, i'd think so. there's new -stable coming, but seems grsec patch has most of the important one commits picked up already 2014-06-26 08:12:08 i got about all musl related bugs fixed that i noticed 2014-06-26 08:12:25 excluding nfs 2014-06-26 08:12:44 yes, we didnt have time for nfs 2014-06-26 08:12:53 i will have a long weekend 2014-06-26 08:12:59 and i want 3.0.1 out asap 2014-06-26 08:13:58 same here 2014-06-26 08:14:45 oh 2014-06-26 08:14:50 there is the issue with xen hvmloader 2014-06-26 08:14:59 i suppose i enable alpine-xen iso 2014-06-26 08:15:05 we ship 3.0.1 with broken xen 2014-06-26 08:15:15 and try fix it for 3.0.2 2014-06-26 08:15:26 i think xen pvs are ok 2014-06-26 08:15:30 only hvmloader is broke 2014-06-26 08:15:42 and in worstcase you can copy hvmloader from v2.7 2014-06-26 08:21:52 i have about 30mins still left before i go 2014-06-26 08:21:59 anything i should look at? 2014-06-26 08:22:45 see if there is anything that looks imporant: http://bugs.alpinelinux.org/projects/alpine/issues?fixed_version_id=82&set_filter=1&status_id=o 2014-06-26 08:22:59 i'm cherry-picking the thigns fixed in edge 2014-06-26 08:23:03 ok. 2014-06-26 08:23:03 xen 2014-06-26 08:23:29 i think we have it all in place 2014-06-26 08:29:02 fabled: rollback of cvs seems to be a low hanging fruit. the build fails though 2014-06-26 08:29:21 you want me to downgrade it? 2014-06-26 08:29:23 ok 2014-06-26 08:29:34 #3053 2014-06-26 08:29:40 http://bugs.alpinelinux.org/issues/3053 2014-06-26 08:30:02 bah, he even posted the needed patch there 2014-06-26 08:30:05 im stupid... 2014-06-26 08:30:06 :) 2014-06-26 08:30:18 https://github.com/sabotage-linux/sabotage/blob/master/KEEP/cvs1.patch 2014-06-26 08:33:09 ncopa, builds with the patch 2014-06-26 08:33:10 yes 2014-06-26 08:33:18 edge+3.0 or 3.0 only? 2014-06-26 08:33:24 both 2014-06-26 08:33:38 should we put the newer feature release in testing? 2014-06-26 08:33:46 nah 2014-06-26 08:34:06 "newer" feature release is like 2009 or so 2014-06-26 08:34:12 i think its dead upstream 2014-06-26 08:34:51 who wants new features for cvs? 2014-06-26 08:35:04 :) 2014-06-26 08:35:06 yeah 2014-06-26 08:35:29 ncopa, i push: http://sprunge.us/IPiV 2014-06-26 08:35:50 looks good 2014-06-26 08:36:01 people will have to apk upgrade -a to actually get it 2014-06-26 08:36:04 yes 2014-06-26 08:36:06 but its ok i think 2014-06-26 08:36:22 ah 2014-06-26 08:36:29 for the cerrypick to 3.0 2014-06-26 08:36:32 git commit ammend 2014-06-26 08:36:45 ugprade -a note? 2014-06-26 08:36:45 and add fixes #3052 2014-06-26 08:36:51 ok 2014-06-26 08:37:01 fixes #3053 2014-06-26 08:38:30 aw 2014-06-26 08:38:38 ? 2014-06-26 08:38:41 i need fix linux-virt-grsec kernel too 2014-06-26 08:53:37 ncopa, full relro bug could be resolved 2014-06-26 08:53:45 it's done for x86 and x86_64 and mostly for arm too 2014-06-26 08:53:50 since arm has no 3.0-builder 2014-06-26 08:53:51 it's ok 2014-06-26 08:53:52 imho 2014-06-26 08:54:15 fabled: its open as a 'set up a 3.0 builder' reminder 2014-06-26 08:54:19 ok 2014-06-26 08:54:30 oh 2014-06-26 08:54:31 reminds me 2014-06-26 08:54:35 there's a bug related to 2014-06-26 08:54:47 mdev 2014-06-26 08:54:59 snd permissions? 2014-06-26 08:55:04 that reminded me 2014-06-26 08:55:05 dahdi 2014-06-26 08:55:40 please file a bug or fix it before you forget again :) 2014-06-26 08:55:45 -dahdi!(.*) root:dialout 0660 =dahdi/%1 2014-06-26 08:55:45 +dahdi/(.*) root:dialout 0660 =dahdi/%1 2014-06-26 08:56:11 that's what i have locally 2014-06-26 08:56:19 hm 2014-06-26 08:56:46 i wonder if that affects csiss or what the hp driver now was called 2014-06-26 08:56:52 hp array 2014-06-26 08:57:21 could be 2014-06-26 08:57:30 seems like the hotplug events changed in recent kernels 2014-06-26 08:57:45 we had similar issue with snd/* 2014-06-26 08:57:46 i need to go now 2014-06-26 08:57:50 no time to file bug. sorry. 2014-06-26 08:57:50 ok see u 2014-06-26 08:58:36 i'll file it 2014-06-26 09:00:24 ncopa: are you ok with splitting libvlc from vlc? 2014-06-26 09:00:51 yes, maybe vlc-libs? 2014-06-26 09:01:02 clandmeter: can I ask why? 2014-06-26 09:01:09 i'm curious why though 2014-06-26 09:01:15 something non-vlc wants libvlc? 2014-06-26 09:01:22 correct 2014-06-26 09:02:04 ubuntu also has libvlc 2014-06-26 09:02:28 fedora doesnt have vlc at all 2014-06-26 09:02:52 atleast not in their own repo 2014-06-26 09:03:19 yeah, you need rpmfusion for multimedia stuff 2014-06-26 09:03:24 I don't even want to know why 2014-06-26 09:05:10 ncopa: you want vlc-libs or libvlc (which is the *official* name) 2014-06-26 09:05:29 vlc-libs 2014-06-26 09:06:02 in face 2014-06-26 09:06:05 incat 2014-06-26 09:06:07 infact 2014-06-26 09:06:19 it should be enough only add that to subpackages 2014-06-26 09:06:36 we have a libs() { default_libs; } split func already 2014-06-26 09:06:44 neat 2014-06-26 09:06:52 wouldnt that spplit also that plugins? 2014-06-26 09:06:59 no 2014-06-26 09:07:10 # predefined splitfunc libs 2014-06-26 09:07:10 default_libs() { 2014-06-26 09:07:11 pkgdesc="$pkgdesc (libraries)" 2014-06-26 09:07:11 local dir= file= 2014-06-26 09:07:11 for dir in lib usr/lib; do 2014-06-26 09:07:11 for file in "$pkgdir"/$dir/lib*.so.[0-9]*; do 2014-06-26 09:07:13 [ -f "$file" ] || continue 2014-06-26 09:07:15 mkdir -p "$subpkgdir"/$dir 2014-06-26 09:07:19 mv "$file" "$subpkgdir"/$dir/ 2014-06-26 09:07:21 done 2014-06-26 09:07:23 done 2014-06-26 09:07:25 } 2014-06-26 09:07:27 but we maybe want the plugins there? 2014-06-26 09:08:11 wow that was fast built... 2014-06-26 09:08:17 yeah maybe 2014-06-26 09:09:16 or split them all together... 2014-06-26 09:09:55 i mean split them individual (not sure that ok) 2014-06-26 09:11:15 dunno 2014-06-26 09:11:32 looks like what ubuntu has 2014-06-26 09:13:31 hmm not all of them 2014-06-26 11:59:26 ncopa: can you cherry pick that commit to 3.0? 2014-06-26 13:24:52 ncopa: sorry about checksum 2014-06-26 13:25:03 the bot doesnt report build failure anymore? 2014-06-26 19:35:07 would be great to have "abort on build failure" again 2014-06-26 19:35:19 and really just ditch old packages from testing… 2014-06-26 20:10:47 ncopa: https://paste.xinu.at/Z0EZ/ list of packages untouched since 2014-03 2014-06-26 20:11:18 ncopa: https://paste.xinu.at/FyI/ and completely non-portable way to generate it 2014-06-26 20:13:13 if we get rid of them, testing will have 162 packages left, reasonable number to review 2014-06-26 20:13:22 in two or three weeks 2014-06-26 20:19:34 oh well, "since" isn't really what I meant, but I guess you got it 2014-06-26 20:19:41 I need more sleep or coffee 2014-06-26 21:32:59 hm, bumping the pkgrel didn't cause a rebuild 2014-06-26 21:34:47 ncopa: any idea why ^^^ ? 2014-06-27 00:07:33 oh man 2014-06-27 00:07:35 this can't be happening 2014-06-27 00:07:43 just restarted alpine 2014-06-27 00:07:59 put in my disk encryption password 2014-06-27 00:08:09 mounting root failed 2014-06-27 00:08:24 initramfs emergency recovery shell launched 2014-06-27 00:08:35 sh: cannot access tty: job control turned off 2014-06-27 00:08:38 sigh 2014-06-27 00:15:22 this is a nightmare 2014-06-27 00:15:46 willing to pay anyone who can help me get it back up 2014-06-27 00:43:19 oh god 2014-06-27 00:43:21 I got it working 2014-06-27 00:43:28 thank god 2014-06-27 00:43:39 I used the instructions here 2014-06-27 00:43:39 http://lists.alpinelinux.org/alpine-devel/3607.html 2014-06-27 00:43:45 they didn't work for me on the first box this happened to 2014-06-27 11:02:53 ncopa: off today? 2014-06-27 11:36:34 heh, he's such a slacker ;) 2014-06-27 11:44:35 barthalion: around? 2014-06-27 11:45:19 hello all! 2014-06-27 11:46:11 hello algitbot 2014-06-27 11:46:52 We made a Kamailio patch... how do you prefer we deal with it? 1 - Apply to AL and send to Kamailio or 2 - Send to Kamailio and wait them to apply the patch? 2014-06-27 11:47:07 it affects AL 2.6 (Kamailio 4.0.4) 2014-06-27 11:47:20 what kind of patch? 2014-06-27 11:48:27 an eye patch 2014-06-27 11:49:11 Well a Silly one: It is just one line but fiz a bug on Polycom Display when making a call. It fixes the NOTIFY method sent to the caller part 2014-06-27 11:49:11 (wow, I'm so glad it's Friday - onm a week day, I would have cried myself to sleep after saying something like that) 2014-06-27 11:49:41 s/fiz/fix/ 2014-06-27 11:49:55 no no, fiz is good 2014-06-27 11:50:11 i guess we can add it to aports, and you can send it to upstream 2014-06-27 11:51:06 clandmeter: thanks a guy called jarruda will send the patch to al-devel and to Kamailio 2014-06-27 11:51:21 ScrumpyJack: It seems you really need to rest :P 2014-06-27 11:51:40 alacerda: np 2014-06-27 11:57:01 isn't 2.6 EOL ? 2014-06-27 11:58:00 ScrumpyJack: it is on 2015 may 2014-06-27 11:58:34 http://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases 2014-06-27 12:03:49 apologies 2014-06-27 12:04:13 ScrumpyJack: that is ok... today is friday 2014-06-27 12:04:16 \o/ 2014-06-27 12:10:03 \o/ 2014-06-27 13:38:04 fabled / ncopa: ping 2014-06-27 13:38:22 any thoughts as to why bumping pkgrel of acf-freeswitch[-vmail] didn't cause rebuild? 2014-06-27 13:38:34 on edge or 3.0 builders 2014-06-27 14:17:29 tdtrask: it fails to build since freeswitch failed to build: http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64/main/acf-freeswitch-vmail/acf-freeswitch-vmail-0.4.0-r0.log 2014-06-27 15:39:37 uggedal: that log is from May 30 2014-06-27 15:40:02 the problem is that freeswitch now builds correctly, but the builder did not try to rebuild the ACF packges 2014-06-27 15:40:14 even when I bumped the pkgrel yesterday 2014-06-27 17:58:04 tdtrask: heh, my bad 2014-06-27 18:07:29 hmm, the edge x86_64 builder seems to be dead 2014-06-27 18:15:56 clandmeter: pong 2014-06-27 18:16:06 uggedal: indeed 2014-06-27 18:16:09 fabled: ↑ 2014-06-28 14:25:08 hi 2014-06-28 14:25:14 finaly i got my apu4c :) 2014-06-28 14:32:23 hi guys 2014-06-28 14:32:36 i cant really find a proper documentation on how to customize a alpine image 2014-06-28 14:32:49 i need to build my own kernel and get a few default applications in there 2014-06-28 14:33:11 can you maybe point me into the right direction? 2014-06-28 16:10:49 you want to make an iso image? 2014-06-28 16:13:21 mkisofs 2014-06-28 16:15:38 it's not about creating my own iso image 2014-06-28 16:15:46 but shipping a different kernel 2014-06-28 16:15:55 (and maybe additional packages) 2014-06-28 16:16:17 i cannot find any info on how you compile/package the kernel for alpine 2014-06-29 12:16:24 azet: everything is in aports 2014-06-29 12:16:38 azet: toolchain, build scripts, etc 2014-06-29 14:47:57 ACTION ran into a problem with kamailio and doesn't see nangel here :( 2014-06-29 14:48:01 # kamdbctl create 2014-06-29 14:48:01 /usr/sbin/kamdbctl: /usr/lib/kamailio//kamctl/kamdbctl.base: line 38: syntax error: missing '}' 2014-06-29 14:51:09 ACTION doesn't see the systax error offhand 2014-06-29 14:55:19 guess I'll create a ticket 2014-06-29 15:00:11 #3104 2014-06-29 15:05:33 barthalion: which packages are that? can't find that anywhere 2014-06-29 15:11:24 azet: mkisofs for customizing images, aports for the rest http://git.alpinelinux.org/cgit/aports/ 2014-06-29 15:15:03 barthalion: for example on debian you get build-essential 2014-06-29 15:15:20 barthalion: i also do not see a kernel package in aports 2014-06-29 15:15:50 barthalion: what i want to do is just compile a custom kernel that is almost identical with the one alpine ships but has a few config changes 2014-06-29 15:15:57 and add apps 2014-06-29 15:16:07 azet: and I gave you the pointers 2014-06-29 15:16:17 linux-grsec is the default kernel in Alpine 2014-06-29 15:16:37 oh k 2014-06-29 15:18:00 counterpart of build-essential is build-base, but alpine-sdk metapackage holds everything you need to make own images 2014-06-29 15:18:07 including build-base 2014-06-29 15:18:10 thanks 2014-06-29 16:38:32 I am trying to install alpinelinux 3.0.1 x86_64 in a vmware workstation while using usb connected SD Card, on first boot i see this line: mdev: bad line 73: 2 tokens found, 3 needed 2014-06-29 16:38:40 http://picpaste.com/pics/XL4RsgiF.1404059828.PNG 2014-06-29 17:38:31 crow: I saw the same line, but the iso booted fine anyway 2014-06-29 17:39:38 tdtrask yes it booted but as i cant boot after installing from this usb i want to report it maybe its related. 2014-06-29 17:40:06 for me right now is hard to install alpinelinux on an headless hardware (apu4c from pcengines) 2014-06-29 18:07:05 is there a way to tell alpine-3.0.1-x86_64.iso burdned on CD to output everything to serial console? (no vga on apu4c)? 2014-06-30 05:46:51 Hi all 2014-06-30 07:16:25 crow: i think you need at the boot prompt type: grsec console=ttyS0 2014-06-30 07:16:26 or so 2014-06-30 07:16:55 or you can extract the files from iso to a temp dir, edit syslinux.cfg and create a new iso 2014-06-30 07:16:59 and burn that 2014-06-30 07:17:40 there was some questions about enabling ttyS0 by default on iso, but you never know what is connected on the serial port... 2014-06-30 07:18:03 well for me it would be nice to have enabled serial as default 2014-06-30 07:18:18 i know it would be nice for some ppl 2014-06-30 07:18:19 well most embeded device have serial for that 2014-06-30 07:18:23 yes 2014-06-30 07:18:24 hm 2014-06-30 07:18:36 and alpinelinux is for "embeded device, small router etc" 2014-06-30 07:19:03 can you boot it from usb? 2014-06-30 07:19:24 ncopa i can boot from usb (did not create bootable usb) 2014-06-30 07:19:26 but i can try 2014-06-30 07:19:32 that would make things easier 2014-06-30 07:19:38 because then you can edit syslinux.cfg 2014-06-30 07:20:01 you can there add serial config for syslinux (so you get boot promt on serial) 2014-06-30 07:20:15 and you can add console=ttyS0 2014-06-30 07:20:21 to default boot line 2014-06-30 07:20:37 yes thats correct 2014-06-30 07:20:45 let me check what i need for make bootable usb 2014-06-30 07:21:00 as alpinelinux iso ist jet not isohybrid 2014-06-30 07:21:12 bootable usb is in theory 2014-06-30 07:21:23 mkfs.vfat 2014-06-30 07:21:33 fdisk 2014-06-30 07:21:46 (and toggle boot partition to be 'bootable') 2014-06-30 07:22:07 then mount it, and use 7zip or soemthing to extract the files 2014-06-30 07:22:14 and finally syslinux 2014-06-30 07:22:32 if you do it from alpine itself we have a tool to do it: setup-bootable 2014-06-30 07:23:05 it should be possible to do it from windows too, format and make it bootable, use 7zip to extract the .iso to usb 2014-06-30 07:23:06 i have an alpinelinux 2.7.8 installation, but does it do latest iso? 2014-06-30 07:23:14 it should 2014-06-30 07:25:33 ncopa: the x86_64 builder seems to be dead 2014-06-30 07:25:52 edge x86_64 that is 2014-06-30 07:26:56 ncopa: and backlog about junk in testing: https://paste.xinu.at/aybowy/ 2014-06-30 07:29:05 ncopa thanks i will try to do it. 2014-06-30 07:30:05 is there anything know, that 3.0.1 does not play very well with the lvm2? when creating volume i got msg about mlock error: no space or something, i wrote these msg but left the pc at home 2014-06-30 07:32:33 uggedal: i'll check it out 2014-06-30 07:33:02 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64.log 2014-06-30 07:33:47 missing dep 2014-06-30 07:33:52 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64/main/umurmur/ 2014-06-30 07:34:12 >>> umurmur: Installing packages on builder: libconfig-dev openssl-dev protobuf-c-dev build-base 2014-06-30 07:34:13 ERROR: unsatisfiable constraints: 2014-06-30 07:34:13 protobuf-c-dev (missing): 2014-06-30 07:34:13 required by: world[protobuf-c-dev] 2014-06-30 07:37:26 seems like build errors are not published here 2014-06-30 07:38:40 crow: i think seth setupdisk script is a bit stupid when it comes to disk usage when the disk is very small 2014-06-30 07:38:51 either that or it is a bug 2014-06-30 07:39:06 well, its likely a bug in any case... 2014-06-30 07:43:47 barthalion: nice 2014-06-30 07:43:57 i doubt i will have time to look at it today 2014-06-30 07:44:28 maybe we should file a bug for a treecleaner project 2014-06-30 07:44:34 file an issue on bugtracker that is 2014-06-30 07:50:58 ncopa actually i did not used setupdisk but manually as i wrote here http://wiki.alpinelinux.org/wiki/Installing_on_GPT_LVM 2014-06-30 07:51:12 i also saw something in dmesg "bia -- "something 2014-06-30 07:53:37 stupid is that i forgot the notebook... 2014-06-30 09:35:00 just wondering how did main/lua-discount got tested and moved to mains' ? 2014-06-30 09:35:26 does it require http://www.pell.portland.or.us/~orc/Code/discount/ ? 2014-06-30 09:37:55 barthalion, re: junk in testing, it also has acf2 ;) 2014-06-30 09:39:50 but would be nice to have things split into manageable groups 2014-06-30 09:44:03 seems nice http://xiki.org 2014-06-30 09:44:55 but requires ruby 2014-06-30 09:46:55 reminds me a bit to zombo.com :) 2014-06-30 09:47:08 "everything is possible" 2014-06-30 11:00:30 nobody felt a need for i/o / perf testing tools in AL 2014-06-30 11:01:36 e.g fio 2014-06-30 11:12:08 what's happened to build-edge-x86_64? Is still down? 2014-06-30 11:13:46 oh 2014-06-30 11:14:28 https://gist.github.com/ncopa/3f898255ec667c8ccf1c 2014-06-30 11:14:35 i think its waf that has hang 2014-06-30 11:15:02 i have seent hat before 2014-06-30 11:15:05 seen* 2014-06-30 11:15:11 dunno why it happens though 2014-06-30 11:15:28 what's waf ? 2014-06-30 11:15:51 build system (like gnu make) implemented in python 2014-06-30 11:16:05 https://code.google.com/p/waf/ 2014-06-30 11:16:30 mpv uses it 2014-06-30 11:16:40 it is likely a parallel issue that is triggered if you have many cores 2014-06-30 11:16:44 i remember stuff hanging with waf earlier too 2014-06-30 11:16:54 x86_64 stopped to work after I pushed new mpv 2014-06-30 11:17:06 not sure if it was waf, or musl issue earlier. 2014-06-30 11:17:11 i killed it 2014-06-30 11:17:11 probably waf issue 2014-06-30 11:17:31 somehow related to pythonn 2014-06-30 11:17:37 some mutex/pthreads relatead issue 2014-06-30 11:17:43 *nod* 2014-06-30 11:17:47 that what i think too 2014-06-30 11:18:34 googling for waf hangs gives some hits 2014-06-30 12:41:44 fyi, armhf builder is busy rebuilding lots of stuff. expect latency until it pushes new things out. 2014-06-30 13:42:54 ncopa: do I need to do anything to get acf-freeswitch[-vmail] to build? I bumped pkgrel, but the builders were down at the time? 2014-06-30 13:46:23 i think they are back now? 2014-06-30 13:52:57 ncopa: but my packages didn't build :( 2014-06-30 13:53:12 build logs are still from May 30 2014-06-30 13:53:54 x86? 2014-06-30 13:53:58 x86_64? 2014-06-30 13:54:01 edge? 2014-06-30 13:54:24 quagga didnt build on edge for some reason... 2014-06-30 13:54:24 ACTION is looking at edge x86_64 2014-06-30 13:54:52 the logging has changed recently 2014-06-30 13:54:59 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64/main/quagga/ 2014-06-30 13:55:05 we do per package logs nowdays 2014-06-30 13:55:40 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64/main/acf-freeswitch/ 2014-06-30 13:56:10 ACTION bumped pkgrel June 26 2014-06-30 13:59:46 b28474a5fe98a99d96393181e9146a2cf9f25df4 2014-06-30 14:01:13 ncopa: thanks, I missed that :) 2014-06-30 14:16:12 ncopa: does a failed build still output to here? 2014-06-30 14:16:36 i thought it did 2014-06-30 14:16:43 oh, no i think it doesnt 2014-06-30 14:16:56 i thinnk it just says upload ready 2014-06-30 14:17:01 because we switched back to don't fail on error 2014-06-30 14:17:01 it doesn't 2014-06-30 14:17:08 you disabled it due to mess in testing 2014-06-30 14:17:15 so it will ignore errors and just move on 2014-06-30 14:17:30 cant you work around that and still display the error? 2014-06-30 14:17:46 not without spamming the channel 2014-06-30 14:17:52 hmm 2014-06-30 14:20:11 i think it should be possible 2014-06-30 14:20:40 we could maybe implement an error hook in abuild 2014-06-30 14:20:44 i think we almost have it already 2014-06-30 14:20:58 call a script on build error 2014-06-30 14:32:58 hm 2014-06-30 14:33:09 i wonder if we should implement buildrepo hooks 2014-06-30 14:34:11 something like /etc/buildrepo/hooks/{pre,post}-build.d 2014-06-30 14:35:17 and for each package it builds it executes the hooks 2014-06-30 14:35:28 we could use that for the mqtt status updates too 2014-06-30 14:38:03 looks that edge is stopped. 2014-06-30 14:38:07 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64.log 2014-06-30 14:38:24 s/edge/build-edge 2014-06-30 14:38:40 it looks like it is done? 2014-06-30 14:38:46 no 2014-06-30 14:38:50 testing built: 0 2014-06-30 14:38:50 testing tried: 143 2014-06-30 14:38:50 testing deleted: 0 2014-06-30 14:38:50 testing total built: 960 2014-06-30 14:38:50 testing total relevant aports: 1103 2014-06-30 14:38:51 testing total aports: 1124 2014-06-30 14:39:02 many things have been skipped 2014-06-30 14:39:06 well 2014-06-30 14:39:10 at least dnscrypt-proxy 2014-06-30 14:39:31 WARNING: dnscrypt-proxy: Skipped due to previous build failure 2014-06-30 14:40:08 this means that the packages has failed or the build-edge has failed? 2014-06-30 14:41:31 it means that dnscrypt-proxy tried to build earlier, but it failed 2014-06-30 14:41:35 then it continued 2014-06-30 14:41:43 ok 2014-06-30 14:41:51 it will not retry build it til the apkbuild is modified 2014-06-30 14:42:13 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64/testing/dnscrypt-proxy/ 2014-06-30 14:42:20 yes, saw it 2014-06-30 14:42:22 checksum error 2014-06-30 18:36:39 ncopa regarding LVM setup i sow last night, here are the logs i get: 2014-06-30 18:36:39 /lib/libdevmapper.so.1.02: mlock failed: Out of memory 2014-06-30 18:36:40 /lib/libdevmapper-event.so.1.02: mlock failed: Out of memory 2014-06-30 18:36:40 dmesg: 2014-06-30 18:36:47 bio: create slab at 1 2014-06-30 18:38:01 Also it seems something is wrong with "network" script or something, on this pc i have two network card (eth0 not connected; eth1 connected). so i gues something went wrong with dns or such. as for example i was not able to setup repository (f to detect fastet) as inet was not working: 2014-06-30 18:38:02 Error this script should be called from udhcpc 2014-06-30 18:38:25 also, when asked about cache : 2014-06-30 18:38:25 Enter apk cache direcotory (or '?' or 'none') [/var/cache/apk] 2014-06-30 18:38:25 ERROR: Package cache is not enabled 2014-06-30 18:38:47 its all after clean boot and start from setup-apline 2014-06-30 19:13:04 something is funky with my local build server ever since I upgraded to musl 2014-06-30 19:13:12 anyone remember where the doc is for the upgrade? 2014-06-30 19:33:07 great :( 2014-06-30 19:33:21 I reinstalled all packages on my build server, and abuild still didn't work 2014-06-30 19:34:59 builder:~/aports/main/acf-kamailio$ abuild fetch 2014-06-30 19:34:59 >>> acf-kamailio: Fetching http://dev.alpinelinux.org/archive/acf-kamailio/acf-kamailio-0.8.1.tar.xz 2014-06-30 19:34:59 Error relocating /usr/lib/libssh2.so.1: fcntl64: symbol not found 2014-06-30 19:34:59 Error relocating /usr/lib/libssh2.so.1: __assert: symbol not found 2014-06-30 19:34:59 Error relocating /usr/lib/libssh2.so.1: __ctype_b: symbol not found 2014-06-30 19:35:00 >>> ERROR: acf-kamailio: fetch failed 2014-06-30 19:35:06 thoughts? 2014-06-30 22:43:03 ncopa also in the vmware workstation and lokal disk i am having some weird msg with lvm 2014-06-30 22:43:03 http://abload.de/img/alpinelinux04f2s9k.png 2014-06-30 22:43:03 http://abload.de/img/alpinelinux05rksap.png 2014-06-30 22:45:42 http://abload.de/img/alpinelinux06husro.png 2014-06-30 23:16:57 ncopa regarding editing syslinux.cfg seems empty i am not sure why.. http://pastebin.com/vnLKRtJM 2014-06-30 23:17:41 i mounted iso to /mnt and did setup-bootable /mnt /dev/sdb2 and then mounted it manually to /media/sdb1 to edit syslinux.cfg