2014-02-01 04:35:58 did you know 2014-02-01 04:36:03 debian-installer 2014-02-01 04:36:06 is the worst thing known to man 2014-02-01 04:36:07 also, hi :p 2014-02-01 10:18:53 Elizacat: lol, hi 2014-02-01 12:18:42 hi 2014-02-01 12:18:55 fabled: you here? 2014-02-01 12:22:56 kaniini, hi 2014-02-01 12:23:04 yeah - back from Milan 2014-02-01 12:24:45 there's a regression in 3.10.27 vs 3.10.23 with mdraid 2014-02-01 12:25:01 /dev/mdXXX is now detected as /dev/mdXXX_0 by mdev 2014-02-01 12:38:32 strange 2014-02-01 12:38:39 kernel is only change? 2014-02-01 19:29:25 ncopa, thnx for this :) http://phoronix.com/forums/showthread.php?94759-X-Org-Server-Systemd-Integration-Proposed/page10&s=aad386b37223185ca102d14afcc43935 2014-02-02 01:33:27 has anyone used the oidentd package? 2014-02-02 03:09:00 yes 2014-02-02 03:09:05 i use it 2014-02-02 03:24:39 do you use it for znc or anything 2014-02-02 03:30:29 kaniini ^ 2014-02-02 03:37:02 no 2014-02-02 03:37:29 oh 2014-02-02 03:37:41 for some reason it always returns NO-USER 2014-02-02 03:37:47 and #znc can't figure it out 2014-02-02 03:37:50 doing everything right :| 2014-02-02 03:38:34 thought it might've been a problem with the package or somethign 2014-02-02 04:10:27 oh. 2014-02-02 04:10:33 you have to run oidentd as root 2014-02-02 04:10:44 or give readproc access 2014-02-02 04:28:15 o 2014-02-02 04:32:32 lol 2014-02-02 04:32:32 ran as root and irc server got an ident response this time 2014-02-02 04:32:32 so uh 2014-02-02 04:32:32 if I don't want to run it as root 2014-02-02 04:32:32 what do I give readproc access to kaniini 2014-02-02 19:28:22 barthalion: around? 2014-02-02 19:29:20 doing some chores, but I can take a break 2014-02-02 19:29:25 clandmeter: ↑ 2014-02-02 19:29:45 what is/are chores? 2014-02-02 19:30:57 making the bed and a dinner at the moment 2014-02-02 19:31:04 barthalion: abi breakage because of libcdio 2014-02-02 19:31:16 hm, thought I rebuilt everything 2014-02-02 19:31:39 what is left? 2014-02-02 19:32:11 audacious plugins 2014-02-02 19:32:22 atleast thats whats broken for me atm 2014-02-02 19:32:40 i can bump it, didnt know if you knew abi changed 2014-02-02 19:32:55 I knew, I even bumped pkgrel 2014-02-02 19:33:07 c528cd344a68 2014-02-02 19:33:32 hmm 2014-02-02 19:33:44 i didnt check git.a.o 2014-02-02 19:34:13 but still the error remains 2014-02-02 19:34:40 feel free to bump it again, maybe I screwed up the order or something 2014-02-02 19:34:43 I'll test it later 2014-02-02 19:34:51 ok ill try to bug it again 2014-02-02 19:35:57 err bump :) 2014-02-02 19:36:45 hmm 2014-02-02 19:37:04 well, yeah, it's surely missing… 2014-02-02 19:37:52 maybe I indeed pushed the rebuild in wrong order 2014-02-02 19:39:19 I thought abuild would take care of the order 2014-02-02 19:39:54 seems to be fine 2014-02-02 19:40:10 I've no idea why audacious-plugins wants previous libcdio so badly 2014-02-02 21:24:28 this one is brilliant 2014-02-02 21:30:26 ncopa__: ↑ 2014-02-03 06:26:35 ncopa, i wonder if we should downgrade abiword: https://bbs.archlinux.org/viewtopic.php?id=176119 2014-02-03 09:04:26 fabled: i pushed abiword fix 2014-02-03 09:31:21 ncopa: does musl repo use different repo keys? 2014-02-03 09:31:32 clandmeter, yes, it has different key 2014-02-03 09:31:42 would be nice if they are added 2014-02-03 09:32:55 fabled: do we have a iso for musl x86_64? 2014-02-03 09:33:12 yes, we have one from sometime ago 2014-02-03 09:33:36 http://dev.alpinelinux.org/~tteras/alpine-140103-x86_64.iso 2014-02-03 09:33:51 i have not tested it, but iirc, someone tested it and it at least boots 2014-02-03 09:34:00 ah ok. thx 2014-02-03 09:34:05 i think i looked there 2014-02-03 09:34:11 but the name didnt ring a bell :) 2014-02-03 09:34:34 could've added musl to it to make it more explicit :) 2014-02-03 09:41:58 fabled: does that iso have the keys added? 2014-02-03 09:42:20 should have, yes 2014-02-03 09:42:46 you added it manually? 2014-02-03 09:45:12 build server adds it's key automatically 2014-02-03 09:46:01 i created a chroot yesterday, but it complained about missing key. 2014-02-03 09:46:08 or am i doing something wrong? 2014-02-03 09:46:27 when doing chroot, you need to add it manually 2014-02-03 09:46:34 for the time being 2014-02-03 11:53:07 ncopa: around? 2014-02-03 12:01:50 hi 2014-02-03 12:04:43 ncopa, libgcrypt does not build on arm due to TEXTRELs, and it seems --disable-asm / --disable-neon-support does not fix it. 2014-02-03 12:05:01 so should we patch out the asm modules from makefile.am? 2014-02-03 12:05:09 fabled: check if gentoo has patch? 2014-02-03 12:05:11 i don't feel like fixing the arm assembly atm 2014-02-03 12:05:25 i'm ok with patch out asm modules from makefile.am 2014-02-03 12:05:33 maybe file a bug upstream 2014-02-03 12:05:43 libgrypt 1.6.x is masked on arm in gentoo 2014-02-03 12:05:57 might be related to your issue 2014-02-03 12:06:23 btw, i dont think apk fetch --progress disable patch solved anything 2014-02-03 12:07:00 try apk fetch --stdout busybox | tar -zt 2014-02-03 12:07:28 hm 2014-02-03 12:08:25 i just dumped with i had in my tree. but yes, the regular logs still mess it up. 2014-02-03 12:08:45 it is right direction though. http://sprunge.us/CNMX 2014-02-03 12:09:07 ok to to push that^^^ 2014-02-03 12:09:08 ? 2014-02-03 12:09:11 i think so 2014-02-03 12:09:15 will not harm anything 2014-02-03 12:09:17 go ahead 2014-02-03 12:09:28 i still can't reproduce the audacity-plugins install_if issue 2014-02-03 12:09:39 it seems to be dependent on multiple things 2014-02-03 12:10:32 i think you need have built audacioius-plugins with old libcdio 2014-02-03 12:10:35 then upgrade it 2014-02-03 12:10:56 so you have an old audacious-plugins in your repo 2014-02-03 12:11:12 then try rebuild new audacious-plugins with new libcdio 2014-02-03 12:12:09 it should be possible to reproduce if you checkout older git and build libcdio 2014-02-03 12:12:22 and link audacious-plugins to that, old libcdio 2014-02-03 12:12:31 then checkout current master HEAD 2014-02-03 12:12:39 and rebild libcdio first 2014-02-03 12:12:41 rebuild* 2014-02-03 12:12:56 and then rebuild audacious-plugins 2014-02-03 12:13:22 it is reproduceable on build-edge servers 2014-02-03 12:44:30 ncopa: is there anyting blocking to add cinnamon? 2014-02-03 12:45:49 clandmeter: its the gtk3 or gtk2 desktop? 2014-02-03 12:46:06 i never remember which is which, cinnamon/mate 2014-02-03 12:46:15 looks like it needs gtk3 2014-02-03 12:48:07 ok 2014-02-03 12:48:14 i dont know any blockers 2014-02-03 12:48:34 and i thin k it might be nice to get rid of the gnome2 stuff... 2014-02-03 12:49:35 ok, then i think ill give it a try and see how far i get with the deps. 2014-02-03 12:51:08 we should dive into testing one day… 2014-02-03 12:51:23 yeah 2014-02-03 12:51:28 lots work needed there 2014-02-03 12:52:26 it will be an evening hobby for me :) 2014-02-03 12:52:54 untill my adruino shipments arrive (if they ever do_ 2014-02-03 12:52:57 ) 2014-02-03 12:53:55 we can also think about kde5, it should be easier to package and split than kde4 2014-02-03 12:54:21 qt5 needs some fixing to 2014-02-03 12:54:25 to build with musl 2014-02-03 12:54:29 and will be surely more modular, no more kdelibs if you want to use just one app 2014-02-03 12:54:40 sounds nice 2014-02-03 12:59:14 ncopa: do you use dropbox on alpine? or is that even possible? 2014-02-03 12:59:47 dropbox is compiled against glibc, so I doubt it 2014-02-03 12:59:50 i dont 2014-02-03 12:59:58 i doubt it it possible yes, due to closed source 2014-02-03 13:00:13 owncloud might be an option 2014-02-03 13:00:27 im thinking of buying a recent chromebook with intel gfx 2014-02-03 13:00:42 but the storage is limted. dropbox could help. 2014-02-03 13:01:19 maybe i should run my own cloud storage. 2014-02-03 13:01:29 with owncloud that is possible yes 2014-02-03 13:01:43 i think mega.co.nz is interesting too 2014-02-03 13:01:52 as i understand it does the crypto on client side 2014-02-03 13:02:04 so they cannot read what you store there (unless they break the crypto) 2014-02-03 13:02:17 Wuala works in the similiar way 2014-02-03 13:02:20 they have a linux client? 2014-02-03 13:03:11 do they have opensource client? 2014-02-03 13:03:23 looks like wuala uses java 2014-02-03 13:03:25 seems itsa java 2014-02-03 13:03:29 :| 2014-02-03 13:03:38 i dont think mega.co.nz has linux client (yet) 2014-02-03 13:07:38 there is megatools 2014-02-03 13:07:48 Dear users, I decided to stop developing megatools, because of uncertainty around the new Mega API license. 2014-02-03 15:45:25 ncopa: were you able to fix the libcdio issue? 2014-02-03 15:46:45 its a problem in apk-tools 2014-02-03 15:46:50 i think fabled looked at it 2014-02-03 15:47:02 but not sure we have a proper sollution for it yet 2014-02-03 15:47:15 i think i can work around it though 2014-02-03 15:47:46 i wanted to install alpine-desktop, but its currently broken (atleast last night) 2014-02-03 15:49:09 x86_64? 2014-02-03 15:49:17 yes 2014-02-03 15:49:53 algitbot: build master 2014-02-03 15:51:00 algitbot: build master 2014-02-03 16:52:57 algitbot: build master 2014-02-03 17:42:04 algitbot: IT IS NOT OKAY 2014-02-03 17:42:08 algitbot: build master 2014-02-03 18:29:15 barthalion: apktools are broken 2014-02-03 18:29:30 I'm so lucky 2014-02-03 18:29:39 everything breaks when I push my changes 2014-02-03 18:29:43 :) 2014-02-03 18:29:55 dont be so pesimistic 2014-02-03 18:30:06 months ago I triggered bug in scripts removing a lot of packages 2014-02-03 18:30:10 now I broke apk-tools :p 2014-02-03 18:30:24 you didnt break it, you found it ;-) 2014-02-04 08:31:06 fabled: any progress with apk-tools' Huh? install_if issue? 2014-02-04 09:10:34 ncopa, no, sorry. need to do few other things first. 2014-02-04 09:10:42 ncopa, seems abiword's save-as-pdf is still broke 2014-02-04 09:14:48 i tried it yesterday and was able to spit out a test.pdf? 2014-02-04 09:14:58 is it x86 uclibc? 2014-02-04 09:15:14 apk version abiword 2014-02-04 09:15:19 does it say -r1? 2014-02-04 09:15:45 build-edge-x86 is still broke due to install_if and apk-tools 2014-02-04 10:19:39 ncopa: it seems that recent gconftool-2 wants to talk to dbus 2014-02-04 10:19:53 i guess thats ok, if it runs... 2014-02-04 10:20:04 heh 2014-02-04 10:20:18 not sure what to do. 2014-02-04 10:20:26 should we igrnore the error? 2014-02-04 10:22:23 i was looking in other places how they manage it. but its hard to understand their logic. 2014-02-04 10:35:59 if we want make 3.0 in May we need fix the remaining musl packages asap 2014-02-04 10:36:54 those are the package missing: https://dpaste.de/Hp0x/raw 2014-02-04 10:37:26 i would say things like ethtool, iputils, kdb, xfsprogs etc are kinda important 2014-02-04 10:37:48 xen would be nice too, but i'd be ok to leave xen out for the 3.0 release if needed 2014-02-04 10:37:49 and what are the failures? or can we easily setup local musl chroot? 2014-02-04 10:38:02 with lxc its simple 2014-02-04 10:39:01 lxc-create -t alpine -n alpine-musl -- --release edge-musl 2014-02-04 10:39:10 might even work from arch 2014-02-04 10:39:12 chroot is working ok, if you got the musl repo key that is. 2014-02-04 10:39:25 ah yes the repo key is needed 2014-02-04 10:39:39 and apk.static I guess 2014-02-04 10:39:45 will take a look in a moment 2014-02-04 10:40:10 what version of lxc is in arch? 2014-02-04 10:40:40 no idea, didn't have time to package it properly 2014-02-04 10:40:48 I guess that either too new or too old 2014-02-04 10:40:55 there was a bunch of patches sent to lxc upstream that should automatically (and securely) download apk.static if needed 2014-02-04 10:41:14 I keep apk.static in my $PATH anyway to recreate chroots 2014-02-04 10:41:33 oh yay, we're up to date 2014-02-04 10:42:04 I'm sure I've seen a rc there though 2014-02-04 10:42:22 then its should probably 'just work' 2014-02-04 10:42:30 hm, seems it's in better state that the last time I've checked 2014-02-04 10:43:09 lxc-create -t alpine ... should work pretty well on all distros, given its new enough 2014-02-04 10:48:29 what was the fix for untrusted signature? 2014-02-04 10:51:22 apparently keys for edge-musl are not there 2014-02-04 11:05:32 barthalion: http://dev.alpinelinux.org/~tteras/alpine-140103-x86_64.iso 2014-02-04 11:05:45 that iso should have tehm 2014-02-04 11:05:58 fabled: where are the apk keys for msul? 2014-02-04 11:06:01 musl* 2014-02-04 11:09:51 ncopa: is there anyway to see the specific build failures for each package? I guess Xen is at least blocked by the Qemu failure, and possibly something else 2014-02-04 11:12:37 some packages are done already, like fish 2014-02-04 11:12:50 clandmeter: checking it now 2014-02-04 11:13:32 barthalion: im not sure though, but fabled provided me that iso as a musl iso. 2014-02-04 11:15:18 I guess it uses the same alpine-keys package 2014-02-04 11:15:40 and I rather feel bad adding --allow-untrusted 2014-02-04 11:16:39 morning 2014-02-04 11:17:34 \o 2014-02-04 12:25:05 royger: ok. I think we have some workaround for qemu 2014-02-04 12:25:15 qemu patches needs to be upstreamed 2014-02-04 14:00:01 hi 2014-02-04 14:00:02 hm... 2014-02-04 14:00:21 apk fix 2014-02-04 14:00:21 (1/1) Reinstalling postfix (2.10.2-r1) 2014-02-04 14:00:21 ERROR: postfix-2.10.2-r1: trying to overwrite usr/sbin/sendmail owned by ssmtp-2.64-r5. 2014-02-04 14:00:21 Executing busybox-1.21.1-r0.trigger 2014-02-04 14:00:21 Executing uclibc-utils-0.9.33.2-r26.trigger 2014-02-04 14:00:22 1 errors; 1791 MiB in 474 packages 2014-02-04 14:01:36 I guess it something to be expected 2014-02-04 14:01:45 postfix and ssmtp provide similiar functionality 2014-02-04 14:01:51 StarWarsFan, you can't have ssmtp and postfix 2014-02-04 14:02:06 ssmtp /or/ postfix 2014-02-04 14:02:07 for the same reason you can have exim and postfix 2014-02-04 14:02:56 hm, the question is, why there are both... 2014-02-04 14:04:26 apk del ssmtp 2014-02-04 14:04:45 yeah, removed postfix 2014-02-04 14:06:48 fabled: i'd like to add this patch to gcc: http://sprunge.us/DBVP 2014-02-04 14:07:13 uh, is it useful? 2014-02-04 14:07:36 you'd need to change all calls to posix_memalign in gcc to __gcc_posix_memalign 2014-02-04 14:08:14 in all c++ files maybe 2014-02-04 14:08:31 but doesn't that alias thing take care of it? 2014-02-04 14:08:41 just makes sure its not exported? 2014-02-04 14:09:07 http://www.openwall.com/lists/musl/2013/11/10/2 2014-02-04 14:11:42 i think the right solution would be to just delete the gcc's declaration 2014-02-04 14:16:28 the surce says: 2014-02-04 14:16:28 may not be visible. */ 2014-02-04 14:16:38 /* We can't depend on since the prototype of posix_memalign 2014-02-04 14:16:38 may not be visible. */ 2014-02-04 14:19:07 the header is broke 2014-02-04 14:19:32 the source file units that use the posix_memalign function, should #define _GNU_SOURCE and #include in them 2014-02-04 14:19:46 not via a wrapper header that tries to declare it self in incompatible way 2014-02-04 14:20:57 the patch you suggests does not fix anything in case posix_memalign is not defined 2014-02-04 14:21:04 alias works the otherway 2014-02-04 14:21:21 extern "C" int __gcc_posix_memalign (void **, size_t, size_t) 2014-02-04 14:21:21 __attribute__((__alias__("posix_memalign"))) throw (); 2014-02-04 14:21:34 defines symbol __gcc_posix_memalign, that when used, will resolve to posix_memalign 2014-02-04 14:21:40 it does not define posix_memalign at all 2014-02-04 14:21:48 ok 2014-02-04 14:22:09 i suppose thats what dalias suggested then 2014-02-04 14:22:12 yes 2014-02-04 14:22:22 he suggested to rename the symbol, and all places where it's used 2014-02-04 14:22:25 use __gcc_posix_memalign in gcc sources 2014-02-04 14:23:00 yeah 2014-02-04 14:33:19 i think the only thing using it is right there 2014-02-04 14:33:24 https://github.com/mirrors/gcc/blob/master/gcc/config/i386/pmm_malloc.h 2014-02-04 14:39:43 so this might work: http://sprunge.us/VOTe 2014-02-04 14:43:08 youd need to change both prototypes 2014-02-04 14:43:15 also the one #ifndef __cplusplus 2014-02-04 14:44:54 oh right 2014-02-04 14:45:00 but without a throw 2014-02-04 14:45:09 right 2014-02-04 14:57:46 the throw should be beofer __attribute__ too i think 2014-02-04 14:57:49 gcc complains.. 2014-02-04 14:57:57 http://pastebin.com/raw.php?i=3kz7UcK4 2014-02-04 15:37:52 algitbot: build 2.7-stable 2014-02-04 15:45:17 files from v2.5.4-243-g410ec3a uploaded 2014-02-04 15:52:11 files from v2.4.11-187-g1e27a08 uploaded 2014-02-04 15:53:07 files from v2.4.11-187-g1e27a08 uploaded 2014-02-04 15:54:31 build-2-5: retry 2014-02-04 19:47:06 hi 2014-02-04 19:47:10 :-9 2014-02-04 19:47:12 :-) 2014-02-04 19:48:42 hi 2014-02-04 19:49:31 how was it? you broke something? :D 2014-02-04 19:50:17 I broke me wife :) 2014-02-04 19:50:26 she was pretty sick when we went home 2014-02-04 19:50:31 some bad food i think 2014-02-04 19:50:42 but for the rest it was nice 2014-02-04 19:50:53 I usually break stuff here 2014-02-04 19:51:16 that is so curious, I had that 2 months ago too - was in holiday and got 1 day sick... 2014-02-04 19:51:29 barthalion: why dont we have pulseaudio? 2014-02-04 19:51:45 but this holiday was really better and great :) 2014-02-04 19:51:51 shafire: she is still sick (alraedy 3 days) 2014-02-04 19:52:04 shafire: so your holidays was good? 2014-02-04 19:52:07 clandmeter: no idea 2014-02-04 19:52:27 cinamon needs pulseaudio 2014-02-04 19:52:30 :| 2014-02-04 19:53:02 aha 2014-02-04 19:54:11 pulseaudio seems like a b*tch 2014-02-04 19:54:58 How is the java status on musl? 2014-02-04 19:55:04 not really 2014-02-04 19:55:07 (java 7) 2014-02-04 19:55:38 barthalion: be my guest :) 2014-02-04 21:09:03 clandmeter: I'll take a look tommorow 2014-02-04 21:28:37 barthalion: im working on it, but having some build errrors with fftw. think ill just leave it out for now. 2014-02-04 22:10:08 is the 3.13 kernel far away (or possible to compile) - LXC 1.0 looks a bit more secure with "unprivileged containers" https://www.stgraber.org/2014/01/17/lxc-1-0-unprivileged-containers/ 2014-02-04 22:21:04 "Kernel: 3.13 + a couple of staging patches (which Ubuntu has in its kernel)" 2014-02-04 22:21:11 3.13 alone won't give much 2014-02-04 22:21:58 3.13 is just a requirement for LXC 1.0 (the alpine repo in 0.90) 2014-02-05 07:31:54 ncopa, new firefox 27. has tls 1.1/1.2 on by default as well as SPDY. will interesting to test it out. 2014-02-05 07:32:42 fabled: yes 2014-02-05 07:32:55 dunno if i will reach to do ff-27 today 2014-02-05 07:33:06 i have a bunch of sec fixes 2014-02-05 07:33:13 and i'd like to do v2.7.4 today too 2014-02-05 07:34:36 ncopa, ok. can you add asterisk-speex to alpine base iso? 2014-02-05 07:35:12 fabled: could you add musl keys to alpine-keys? 2014-02-05 07:35:16 it was split out earlier, and now i have some .iso only installs that fail due to certain asterisk modules missing 2014-02-05 07:35:36 barthalion, mmm. global, or to the musl build's alpine-keys pkg? 2014-02-05 07:36:06 probably to the musl build, it really doesn't matter to me 2014-02-05 07:36:08 fabled: http://bugs.alpinelinux.org/issues/2676 2014-02-05 07:36:17 just having the keys provided somewhere 2014-02-05 07:36:33 somwhere easy to find... 2014-02-05 07:37:23 the iso has them on the initramfs, and it installs them to new installs. but yes, having them in alpine-keys would be good 2014-02-05 07:37:37 i just wonder if we want them globally added, or just musl variant. 2014-02-05 07:37:55 ncopa, any preference? 2014-02-05 07:38:06 i'm ok global 2014-02-05 07:38:13 i thin i'd prefer global 2014-02-05 07:39:43 ok 2014-02-05 07:39:48 so i just push them to alpine-keys? 2014-02-05 07:40:24 yeah 2014-02-05 07:42:53 urgh 2014-02-05 07:42:54 checksum 2014-02-05 07:43:33 is it only me who actually builds the packages before pushing ;) 2014-02-05 07:43:49 oh, audacious is still broke. 2014-02-05 07:43:54 the install_if issue 2014-02-05 07:43:57 yup 2014-02-05 07:44:15 i manually cleaned the old pkgs in edge x86_64 2014-02-05 07:44:23 ncopa, i generally build things first. in coffee deprived state i might skip it... ;) 2014-02-05 07:44:35 ncopa: I build packages locally too if I'm not on battery 2014-02-05 07:45:08 the next person who pushes checksumfailure will get the job to write a git hook 2014-02-05 07:45:19 haha 2014-02-05 07:45:29 that does the checksum check and rejects the git push (or reverts it) 2014-02-05 07:45:34 rejects 2014-02-05 07:45:43 it should be local commit hook i think 2014-02-05 07:45:54 then it need to download things while you waith for the git push to complete 2014-02-05 07:46:06 it *should* be a local commit hook 2014-02-05 07:46:17 but the git server cannot enforce that 2014-02-05 07:46:39 we can remove push keys for a week as punishment for not using the hook 2014-02-05 07:46:56 for not testing if package builds 2014-02-05 07:47:18 i'd say the opposite, remove all keys *except* the person who mess up 2014-02-05 07:47:25 so he'll have to do it all 2014-02-05 07:47:27 :D 2014-02-05 07:47:28 for a week 2014-02-05 07:48:59 kaniini, clandmeter: did you get that? ^^^ 2014-02-05 07:49:24 no im not here 2014-02-05 07:49:37 next person who git pushes somethign that fails pass abuild checksum test will have to write a git hook script 2014-02-05 07:49:45 hi 2014-02-05 07:50:00 got it 2014-02-05 07:50:00 deal? 2014-02-05 07:50:03 good :) 2014-02-05 07:50:16 rnalrd: that applies to you too ^^^ 2014-02-05 07:50:34 ill wait till the first one breaks it. 2014-02-05 07:50:51 :) 2014-02-05 07:51:34 i have enough changes in my local branch to break whole aports, instead of just checksums :) 2014-02-05 07:51:46 ncopa: i htink i have speed up gconf 2014-02-05 07:51:59 urgh lol 2014-02-05 07:52:27 clandmeter: cool! 2014-02-05 07:52:44 I don't have to have changes breaking whole aports, I do it anyway 2014-02-05 07:53:03 it's a real gift 2014-02-05 07:53:43 doing it properly on serverside is slightly tricky 2014-02-05 07:54:04 but i'd be ok if everyone has a local git hook that checks it pre-commit 2014-02-05 09:40:03 clandmeter: try fftw-3.3.3-r1 ^^^ 2014-02-05 09:40:04 ok 2014-02-05 09:51:03 ncopa: works! 2014-02-05 09:51:15 good! 2014-02-05 09:51:16 thanks! 2014-02-05 09:52:06 ncopa: are you ok, if i push untested pkgs into testing? 2014-02-05 09:52:45 as long as they builds yes 2014-02-05 09:53:04 then i can pull them in on a machine that can actually run X 2014-02-05 09:53:10 the point with 'tesing' is to push untested things to there, so you can test them 2014-02-05 09:53:21 so thats perfectly ok 2014-02-05 09:53:41 you can look that they look ok, that they dont install files in /usr/local and such 2014-02-05 09:53:59 but you dont need to actually testrun before push to testing 2014-02-05 11:45:37 ncopa: would you move quassel from testing to main? 2014-02-05 12:05:39 StarWarsFan: sure. I'd be happy to 2014-02-05 12:06:11 WARNING: Total of 56 unsatisfiable package names. Your repository maybe broken. 2014-02-05 12:07:22 barthalion: what does apk dot --errors say? 2014-02-05 12:07:28 might give a hint on whats broke 2014-02-05 12:10:16 ncopa: do we still have a central point for build logs? 2014-02-05 12:10:38 yes and no 2014-02-05 12:10:45 we have build logs per arch 2014-02-05 12:11:13 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64.log 2014-02-05 12:11:23 http://bld1.alpinelinux.org/buildlogs/build-edge-x86.log 2014-02-05 12:11:32 would be nice if algitbot would say an url to a tailing log file. 2014-02-05 12:11:39 when it starts to build. 2014-02-05 12:11:52 dont you think its noisy as it is? 2014-02-05 12:12:06 ncopa: would be great 2014-02-05 12:12:12 i'd like to stay as maintainer 2014-02-05 12:12:26 i think functionaly is more important. 2014-02-05 12:12:30 or have a central point 2014-02-05 12:12:34 and add it to the topic 2014-02-05 12:13:08 we have multiple build servers so central point needs some work 2014-02-05 12:13:18 but is possible with nfs for example 2014-02-05 12:13:18 i have no idea if the bot is building now 2014-02-05 12:13:42 well, we could have log.a.o 2014-02-05 12:13:47 like we have msg 2014-02-05 12:13:59 or just use msg 2014-02-05 12:14:16 the two url's above should be the the logs 2014-02-05 12:14:45 i'll think of it 2014-02-05 12:14:50 but not tailing? 2014-02-05 12:14:55 but nno chance to implement it today 2014-02-05 12:15:07 its over http 2014-02-05 12:15:12 how do you tail over http? 2014-02-05 12:15:19 like i did before 2014-02-05 12:15:22 js-logtail 2014-02-05 12:15:36 but you keep breaking it 2014-02-05 12:15:39 ;) 2014-02-05 12:15:46 i got new build servers... 2014-02-05 12:16:04 faster building or keep old logtail... choice was simple.. 2014-02-05 12:16:06 reminds me 2014-02-05 12:16:15 i still have some old boxes on the old build.a.o 2014-02-05 12:16:16 we can have logs.a.o which has links to the build servers 2014-02-05 12:16:17 sigh... 2014-02-05 12:16:25 yeah 2014-02-05 12:16:42 then we just add that to topic, untill everybody got used to it 2014-02-05 12:16:53 i think what we should do is have a http revproxy on bld1 and bld2 2014-02-05 12:17:15 will think of it 2014-02-05 12:17:40 i had plan to do nfs for distfiles and build logs 2014-02-05 12:17:46 so the build servers share distfiles 2014-02-05 12:24:57 btw 2014-02-05 12:25:12 clandmeter: what are you interested in re build servers? 2014-02-05 12:25:14 the build status? 2014-02-05 12:25:20 which package it is building? 2014-02-05 12:25:27 if its building or idle? 2014-02-05 12:25:31 we want to see what's happing in the exact moment 2014-02-05 12:25:48 or do you want the exact build log output? 2014-02-05 12:26:27 because now we can't know if buildbot went maverick or it will throw a error after hour or two 2014-02-05 12:26:30 i think the buildlog provides all that information. 2014-02-05 12:26:59 but its mainly the current build server status that is mostyl interesting right? 2014-02-05 12:27:10 but remembering where its at can be hard 2014-02-05 12:27:43 for now you can use this: mosquitto_sub -v -h msg.alpinelinux.org -t 'build/+' 2014-02-05 12:27:58 :) 2014-02-05 12:28:13 it will tell what it is building 2014-02-05 12:28:15 or if its idle 2014-02-05 12:28:30 the intention was to make a webgui for that 2014-02-05 12:29:48 what about x86? 2014-02-05 12:29:55 is that machine building? 2014-02-05 12:30:04 build/build-edge-x86 building main/bind-9.9.5-r0 2014-02-05 12:30:04 build/build-edge-musl-x86 ERROR: dvgrab: Aborted by user 2014-02-05 12:30:05 build/build-edge-x86_64 idle 2014-02-05 12:30:47 seems to work 2014-02-05 12:30:53 what if one of the machines is down? 2014-02-05 12:31:56 it shows the last message 2014-02-05 12:32:04 they publish them as 'retained' messages 2014-02-05 12:32:34 if you diconnect and reconnect to mqtt, it will show the last published message 2014-02-05 12:32:39 just like a log file does.. 2014-02-05 12:33:13 the another issue I've had is that I couldn't get 2.7-stable builder to actually do the build 2014-02-05 12:34:13 ah x86 is fix now? 2014-02-05 12:34:29 clandmeter: i kickstarted it 2014-02-05 12:34:38 as fabled dont have time to fix the underlying issue atm 2014-02-05 12:35:03 barthalion: 2.7-stable builders are broken and needs babysitting 2014-02-05 12:35:12 they use the zmq based script i did 2014-02-05 12:35:25 if i get time I'll try replace those with the mqtt based today 2014-02-05 12:36:34 ah, I thought you switched them already 2014-02-05 12:37:08 it explains everything then 2014-02-05 12:40:36 sorry i havent 2014-02-05 12:40:40 and it scares me 2014-02-05 12:40:59 because i'm not sure it will be able to create a release when i tag it 2014-02-05 12:41:04 til i actually tag it... 2014-02-05 12:41:51 well, no problem, I just turns me off a bit 2014-02-05 12:42:26 i do have problems with keeping up with everything 2014-02-05 12:42:31 in fact you never know if packages you pushed will be built 2014-02-05 12:42:46 i'll try prioritize it... 2014-02-05 12:44:21 hm, can I make mosquitto_sub quit immadietly after shwoing last status? 2014-02-05 12:44:43 doesn't seem to 2014-02-05 12:44:54 I can live without it 2014-02-05 13:05:02 clandmeter: alacerda wants add a "follow alpine on twitter" on www.a.o 2014-02-05 13:05:08 can you please help him with that? 2014-02-05 13:06:20 yeah ill help him 2014-02-05 13:06:30 thanks ncopa and clandmeter ;) 2014-02-05 13:06:40 thanks to both of you too :) 2014-02-05 13:07:43 fabled: did you test the strongswan init.d script? 2014-02-05 13:08:17 do you think we could move strongswan to main, backport it to v2.7 and close http://bugs.alpinelinux.org/issues/1852 ? 2014-02-05 13:09:59 ncopa, i've tested it to run and work... but not the init.d script. 2014-02-05 13:10:43 you started it with 'ipsec start' ? 2014-02-05 13:11:17 Mp5shooter: it was you who wanted libreswan right? 2014-02-05 13:11:22 can you use strongswan instead? 2014-02-05 13:21:04 ACTION says: If someone wants to make a video teaching how to do something on Alpine Linux. Record the video and send the video link to social@al.o 2014-02-05 13:42:10 alacerda: there was this guy on youtube that kind of tried to do something with alpine (whatever that was) 2014-02-05 13:43:19 who? 2014-02-05 13:43:58 also, if someone want some bacgrkound music (simple piano chords) let me know 2014-02-05 13:44:06 and i might be able to make something 2014-02-05 13:44:28 ncopa: do you play piano? :| 2014-02-05 13:44:34 sometimes... 2014-02-05 13:44:44 =] congrats 2014-02-05 13:54:20 rebooting. brb (hopefully) 2014-02-05 14:05:43 ncopa: do you sing also? 2014-02-05 14:06:17 clandmeter: only backing vocal 2014-02-05 14:32:55 wow 2014-02-05 15:23:15 hi 2014-02-05 15:23:23 what is the current musl status?:) 2014-02-05 15:27:33 iirc there was 69 missing packages last time i tried 2014-02-05 15:27:37 for x86_64 2014-02-05 15:27:53 i run musl on my private laptop 2014-02-05 15:29:38 sounds very good 2014-02-05 15:30:24 what I want to know, can I mount tmpfs with noatime? 2014-02-05 15:30:51 dunno 2014-02-05 15:30:56 why would you? 2014-02-05 15:31:37 I read that it removes a write layer 2014-02-05 15:32:02 i am switching the v2.7 builders to mqtt based messaging 2014-02-05 15:32:06 should make it more stable 2014-02-05 15:39:03 ncopa: could you share a up-to-date missing package list? 2014-02-05 15:39:20 soon 2014-02-05 15:40:40 barthalion: v2.7 builservers now uses mqtt 2014-02-05 15:40:47 this is first release with that 2014-02-05 15:40:52 i hope nothing breaks... 2014-02-05 15:42:07 hm 2014-02-05 15:42:15 it didnt pick up the new release 2014-02-05 15:42:22 just a quick, what have you used before mqtt and why the change? 2014-02-05 15:42:39 homebuilt zmq scripts 2014-02-05 15:42:44 it was horribly unstable 2014-02-05 15:43:29 zeromq? 2014-02-05 15:43:34 yes 2014-02-05 15:44:11 good to know, ok 2014-02-05 15:44:41 sweet! 2014-02-05 15:44:46 it *did* pick up the release 2014-02-05 15:44:56 the upload just did not get logged 2014-02-05 15:46:14 oh, new release 2014-02-05 15:47:03 \0/ new release 2014-02-05 15:48:50 http://alpinelinux.org/release-2.7.4 2014-02-05 15:50:19 ncopa: great 2014-02-05 15:51:38 ACTION is ready to download it 2014-02-05 15:56:37 main total built: 1659 2014-02-05 15:56:38 main total relevant aports: 1729 2014-02-05 15:56:42 thats musl 2014-02-05 15:56:51 testing total relevant aports: 1093 2014-02-05 15:56:51 testing total aports: 1115 2014-02-05 15:57:00 lots of packages in testing also just builds 2014-02-05 15:57:49 barthalion: this is the not-yet-built list for musl: http://sprunge.us/FcPS 2014-02-05 15:58:20 sure it's up to date? 2014-02-05 15:58:28 fish shouldn't be there 2014-02-05 15:58:54 or is it for x86_64? 2014-02-05 15:59:14 note that it does not list packages that are not tried due to a dep failed 2014-02-05 15:59:16 its x86_64 2014-02-05 16:00:11 are builders for both archs up? 2014-02-05 16:00:18 msul? 2014-02-05 16:00:18 no 2014-02-05 16:00:21 not auto 2014-02-05 16:00:37 ah, ok 2014-02-05 16:00:38 i didnt enable the autobuilder yet because we wanted it continue on failure 2014-02-05 16:00:47 so i have rewritten the buildrepo script in lua 2014-02-05 16:00:53 its lot faster now 2014-02-05 16:01:21 but i need connect it with the current build scripts 2014-02-05 16:01:32 didnt want to do that til i had migrated aports-build to v2.7 2014-02-05 16:01:54 fish buildlog: http://sprunge.us/iKDb 2014-02-05 16:04:15 oh did i mention that new buildrepo.lua generates per-package logs? 2014-02-05 16:04:18 gotta go 2014-02-05 16:04:19 see u 2014-02-05 16:07:08 on abuild -r: 2014-02-05 16:07:08 WARNING: Ignoring /home/build/packages//build/x86_64/APKINDEX.tar.gz: No such file or directory 2014-02-05 16:07:11 ERROR: You tried to add a non-repository package to system, but it would be lost on next reboot. Enable package caching (apk cache --help) or use --force if you know what you are doing. 2014-02-05 16:08:41 no buildchroot for me :( 2014-02-05 16:14:44 http://alpinelinux.org/content/alpine-linux-gallery 2014-02-05 16:22:39 fabled, ncopa: is there any constant in musl I can use for #if defined()? 2014-02-05 16:23:01 hey ncopa, yes I was the one who wanted libreswan 2014-02-05 16:23:03 oh wait, I guess no 2014-02-05 16:23:21 let me guess, there is no such thing because it's ugly from the definition 2014-02-05 16:35:30 yay, fish fixed 2014-02-05 16:41:23 now openjdk7 :op 2014-02-05 16:41:47 I'm starting small, sorry 2014-02-05 16:43:09 :-) 2014-02-05 16:43:42 flashcache-utils? :) 2014-02-05 16:43:59 working on fsarchiver 2014-02-05 16:51:21 well, a bit too much for me 2014-02-05 16:51:27 maybe tommorow 2014-02-05 21:18:10 barthalion: when are you going to start on libre office? 2014-02-05 21:31:32 was I going? 2014-02-05 21:32:00 well, if I was, probably after pushing pulseaudio and fixing some musl errors 2014-02-05 21:40:27 barthalion: are you working on pulsadio 2014-02-05 21:40:30 ? 2014-02-05 21:41:00 kind of 2014-02-05 21:41:16 i have it done here 2014-02-05 21:41:18 :) 2014-02-05 21:41:34 good, one thing less from todo :) 2014-02-05 21:41:46 including a bunch of other 2014-02-05 21:41:49 Hi. Can someone help me with building a custom iso image? I'm stuck at WARNING: File '/root/alpine-iso/isotmp.alpine-mini/isofs/apks/x86_64/*.apk' is unaccessible. I'm using the alpine-mini configuration. 2014-02-05 21:41:49 hmm 2014-02-05 21:41:50 for cinnamon 2014-02-05 21:42:00 anyone want to package firefox sync? 2014-02-05 21:43:09 isn't it a firefox feature? 2014-02-05 21:46:25 they hava firefox sync server available if you want to host your own 2014-02-05 21:47:13 plueschy: that error doesnt ring a bell 2014-02-05 21:47:37 did you modify the profile? 2014-02-05 21:50:56 barthalion: i thought you would do libre office after you had done openjdk 2014-02-05 21:51:36 clandmeter: I think I've fixed it now. Had to remove the line "/media/cdrom/apks" in /etc/apk/repositories 2014-02-05 21:51:47 I'm playing around small musl related compilation errors now, I guess that ncopa can throw 13901 fixes per minute 2014-02-05 21:52:19 I'm not so C-skilled that some other devs here are 2014-02-05 21:52:53 barthalion: :) tell me about it. 2014-02-05 21:53:13 i can just spell b.a.s.h. 2014-02-05 21:53:28 just like me 2014-02-05 21:53:35 and even not always correctly 2014-02-05 21:53:41 or using all it features :p 2014-02-05 21:54:10 bb shell has no features :p 2014-02-05 21:55:02 http://blog.docker.io/2014/02/docker-0-8-quality-new-builder-features-btrfs-storage-osx-support/ guess who's on the contributors list 2014-02-05 21:55:33 barthalion: you? 2014-02-05 21:55:38 yep 2014-02-05 21:55:41 gz :3 2014-02-05 21:55:50 small changes unfortunately, not directly related to project 2014-02-05 21:56:11 I'm going to push script for Alpine image creation yet 2014-02-05 21:56:23 I mostly fixed Arch one this time 2014-02-05 21:59:20 I have to modify the Makefile to make changes on the iso rootfs, right? 2014-02-05 21:59:31 I think so 2014-02-05 21:59:35 okay... 2014-02-05 22:00:06 I'm not a good person to ask, I haven't played with making iso yet 2014-02-05 22:00:20 you should wait for ncopa if you encounter any issues 2014-02-05 22:00:24 or ask on alpine-devel 2014-02-05 22:00:33 (I mean mailing list) 2014-02-05 22:01:12 it all depends what you want to change i guess. 2014-02-05 22:01:31 but like barthalion says, ncopa is your guy. 2014-02-05 22:01:50 I'm trying to build an iso that will directly configure network via dhcp and start an openssh server. Will automate the installation with ansible then. 2014-02-05 22:02:01 ah 2014-02-05 22:02:15 then thats not needed 2014-02-05 22:02:22 you can add your own overlay 2014-02-05 22:02:28 oh 2014-02-05 22:02:38 boot the cd 2014-02-05 22:02:42 make your changes 2014-02-05 22:02:45 lbu_commit 2014-02-05 22:03:05 im not sure how to add it though 2014-02-05 22:03:13 never did it mysefl 2014-02-05 22:03:45 lbu_commit will create ovl file (i guess you know this part) 2014-02-05 22:04:07 no i don't ... 2014-02-05 22:04:13 lbu is new for me ^^ 2014-02-05 22:04:24 I'll check out the wiki i guess 2014-02-05 22:04:36 lbu manages your changes on a memory based system 2014-02-05 22:04:49 by default it will monitor /etc 2014-02-05 22:05:00 but you can add other things with lbu_include 2014-02-05 22:05:05 okay 2014-02-05 22:05:19 its all on the wiki afaik 2014-02-05 22:05:26 sounds good :) ty clandmeter 2014-02-05 22:05:57 jbilyk_: didnt you include overlay in iso's? 2014-02-05 23:10:33 barthalion: ^ 2014-02-06 02:28:18 so 2014-02-06 02:28:22 is libreswan broken? 2014-02-06 06:44:00 How to create a package for current stable release ? Do we have separate repo for it ? 2014-02-06 06:45:37 what about ld.so.conf.d directory ? how to add ld.so.conf directory from a package ? 2014-02-06 08:04:48 KozakIvan: new packages always go to edge first. 2014-02-06 08:29:57 KozakIvan: you can add a file to ld.so.conf.d 2014-02-06 08:52:26 barthalion: do you know how to prevent docbook errors? 2014-02-06 08:52:42 trying to download schema from sf.net, but fails... 2014-02-06 08:52:48 i think ive seen that before. 2014-02-06 08:57:09 found it. 2014-02-06 10:04:08 hi 2014-02-06 10:20:14 hi 2014-02-06 10:22:41 Hey Hey 2014-02-06 10:22:43 what is going on? 2014-02-06 10:22:45 fse 2014-02-06 10:22:53 gefe 2014-02-06 10:22:59 rggerg 2014-02-06 10:23:01 ger 2014-02-06 10:23:03 ger 2014-02-06 10:23:05 ger 2014-02-06 10:23:07 gre 2014-02-06 10:23:09 g 2014-02-06 10:23:11 reg 2014-02-06 10:23:13 er 2014-02-06 10:23:15 ger 2014-02-06 10:23:17 g 2014-02-06 10:23:21 reg 2014-02-06 10:23:22 ;-) 2014-02-06 10:23:23 reg 2014-02-06 10:23:25 rger 2014-02-06 10:23:27 er 2014-02-06 10:23:29 ger 2014-02-06 10:23:31 ger 2014-02-06 10:23:33 clandmeter: ^^ 2014-02-06 10:23:33 ger 2014-02-06 10:23:35 ge 2014-02-06 10:23:37 rg 2014-02-06 10:23:39 erg 2014-02-06 10:23:41 erg 2014-02-06 10:23:42 alex__: yeah, keep spamming 2014-02-06 10:23:43 e 2014-02-06 10:23:45 ger 2014-02-06 10:23:47 excellent idea 2014-02-06 10:23:47 ge 2014-02-06 10:23:50 haha 2014-02-06 10:23:51 rg 2014-02-06 10:23:53 erg 2014-02-06 10:23:55 eg 2014-02-06 10:23:57 er 2014-02-06 10:23:59 ge 2014-02-06 10:24:01 rge 2014-02-06 10:24:03 rg 2014-02-06 10:24:05 er 2014-02-06 10:24:06 God bless /ignore 2014-02-06 10:24:07 ger 2014-02-06 10:24:09 ger 2014-02-06 10:24:11 best day today (so much activity in this channel) 2014-02-06 10:24:11 ge 2014-02-06 10:24:11 Sup 2014-02-06 10:24:13 rger 2014-02-06 10:24:15 ger 2014-02-06 10:24:17 how the hell do i ban somebody? 2014-02-06 10:24:17 g 2014-02-06 10:24:21 erg 2014-02-06 10:24:23 er 2014-02-06 10:24:24 clandmeter: with /ban :p 2014-02-06 10:24:25 ge 2014-02-06 10:24:27 rg 2014-02-06 10:24:29 erg 2014-02-06 10:24:31 er 2014-02-06 10:24:33 Dont know, ask the internet 2014-02-06 10:24:33 g 2014-02-06 10:24:35 erg 2014-02-06 10:24:37 er 2014-02-06 10:24:41 Svenblodthoerst == alex__ :) 2014-02-06 10:24:48 yep 2014-02-06 10:24:54 Hahaha Alex ! 2014-02-06 10:25:17 Whats up?! 2014-02-06 10:25:25 clandmeter: you may want to ban Svenblodthoerst as well, the same host 2014-02-06 10:26:02 set a ban on *@p4FF10FFC.dip0.t-ipconnect.de 2014-02-06 10:26:45 so if i ban him he cant say anything anymore? 2014-02-06 10:26:52 he can't join 2014-02-06 10:27:07 thats not my question :) 2014-02-06 10:27:17 he cant say anything anymore 2014-02-06 10:27:19 :p 2014-02-06 10:27:25 you should in fact use /kickban 2014-02-06 10:27:26 cause when i banned him he stopped 2014-02-06 10:34:55 he spammed a lot more channels... 2014-02-06 12:00:04 ncopa: around? 2014-02-06 12:03:09 oh gtk3 has gtk2 as direct dep. 2014-02-06 12:08:48 clandmeter: ncopa told me that today he would travel 2014-02-06 12:09:10 ok thx alacerda 2014-02-06 12:09:25 i already had that idea he was not around 2014-02-06 15:28:04 barthalion: not going to break anything today? 2014-02-06 15:45:46 clandmeter: sorry, not today 2014-02-07 09:42:52 hi 2014-02-07 11:11:29 oh yay, breakage 2014-02-07 11:11:38 imagemagick bumped soname in 6.8.8.2 2014-02-07 11:13:11 yay, moar breakage 2014-02-07 11:29:56 let's see if I can fix it later 2014-02-07 11:31:04 barthalion: do you now how to change boot kernel "the alpine way"? any script for that? 2014-02-07 11:33:02 ncopa mentioned some script, but guess it's still in the works 2014-02-07 11:33:14 i thought it would add the boot option automatically when install a different kernel... 2014-02-07 12:08:02 everything builds fine when I've a lot of time and fails when I need to go 2014-02-07 12:08:17 it's new Murphy's law 2014-02-07 12:13:00 have someone of you a registered trademark? 2014-02-07 12:13:09 does ... have* 2014-02-07 16:36:16 With 3 LXC containers on a single host (mysql / php / nginx) - is openvswitch overkill or more efficient than the bridging built into the kernel ? 2014-02-07 17:36:08 test 2014-02-08 10:03:48 BitL0G1c: i would say normal linux bridging is fine 2014-02-08 13:00:31 ok, inkscape fixed 2014-02-08 13:00:43 qt5 left 2014-02-08 13:19:52 seems I could wait with inkscape 2014-02-08 13:20:07 yet another rebuild when I push new boost… 2014-02-08 13:21:42 argh 2014-02-08 13:26:46 3 weeks later 2014-02-08 15:14:19 hi 2014-02-08 15:14:44 kaniini: do you have alpine 2.7 openvz templates? 2014-02-08 15:17:45 no 2014-02-09 01:12:26 so 2014-02-09 01:12:28 is libreswan broken? 2014-02-09 13:27:23 barthalion, you around 2014-02-09 13:27:26 ? 2014-02-09 13:28:21 could you push testing/php-zend-opcache/ to unmaintained ? 2014-02-09 13:29:05 main/php 5.5.+ has opcache build into 2014-02-09 13:29:57 sure, will do in a moment 2014-02-09 13:30:05 I'll actually remove it 2014-02-09 13:30:35 ok 2014-02-09 13:30:50 thanks 2014-02-09 23:11:02 is there any reason why MariaDB is not in the repositories ? 2014-02-09 23:13:57 because we're still on MySQL 2014-02-09 23:15:48 mainly because we (as in Arch) encountered some problems during migration 2014-02-09 23:16:03 ending up with requirement of user manual intervention 2014-02-09 23:16:34 ah ok 2014-02-09 23:16:34 if ncopa won't change his mind, we wait for stable mariadb 10 2014-02-09 23:16:48 it will require manual intervention anyway 2014-02-09 23:17:19 but it's much better excuse than "ohai, run 3 commands or your database may not work" 2014-02-09 23:18:57 (it's better because upgrade to mysql 5.6 would require the same steps, so…) 2014-02-09 23:21:42 ...wait for 10.1 stable 2014-02-09 23:24:23 well, we're talking about Alpine, we'll obviously use what will work flawlessly :p 2014-02-09 23:26:22 LXC works great - finally got networking working in the containers 2014-02-10 12:09:09 hey, you got cinnamon working? 2014-02-10 12:09:33 clandmeter: ^^^ 2014-02-10 12:15:49 ncopa: not yet 2014-02-10 12:16:06 not sure whats wrong, trying to find out. 2014-02-10 12:22:53 but it builds? 2014-02-10 12:22:55 very nice 2014-02-10 12:27:02 yes it does 2014-02-10 12:27:10 actually pretty easy to build 2014-02-10 12:27:25 not much patching (if any at all) 2014-02-10 13:09:03 ncopa: do i actually need to compile the glib schemas per package, or doest the glib trigger handle it? 2014-02-10 13:11:54 i think the glib trigger should handle it 2014-02-10 21:06:44 I'm trying to run "apk fix --root $MNT linux-grsec" from a CD to repair a broken initramfs (missing cryptsetup modules) - EXTLINUX complains it cannot open "/dev/vda1" - does "setup-disk" install some packages not in the CD image ? 2014-02-10 21:34:23 hi 2014-02-10 21:34:33 how can I install the gnu "ps"? 2014-02-10 21:34:56 sry, wrong channel 2014-02-10 23:38:53 building vmlinuz & initramfs on a working machine allowed me to fix it - will raise a bug & patch to fix this for LUKS installs 2014-02-10 23:40:20 oh man 2014-02-10 23:40:28 i hope debian actually switches to systemd by default 2014-02-10 23:40:36 that will be the best thing to have ever happened to alpine 2014-02-10 23:40:39 lol 2014-02-10 23:40:50 why? 2014-02-10 23:41:31 because people working on debian will go look for debian-like alternatives 2014-02-10 23:41:36 amongst them, alpine 2014-02-10 23:53:16 :-) 2014-02-10 23:53:17 good night 2014-02-11 21:40:17 do installations on LVM run without any initramfs problems when the kernel is upgraded ? (I'm patching mkinitramfs.trigger for LUKS / cryptsetup - & wonder if it's a problem on LVM too ?) 2014-02-12 01:13:44 so 2014-02-12 01:13:47 is libreswan broken? 2014-02-12 05:51:06 does php-fpm always show as crashed in rc-status for anyone else 2014-02-12 05:51:11 even when it's running and working fine? 2014-02-12 07:24:42 BitL0G1c: I use LVM here and it appears to work fine 2014-02-12 07:25:10 Mp5shooter: i dont think libreswan is broken. but i havent tested so i dont really know 2014-02-12 07:25:34 Mp5shooter: can you please file a bug for php-fpm? thanks 2014-02-12 09:11:08 ncopa: do you need more info from me for the move of quassel from testing to main? 2014-02-12 09:11:15 or should i file a ticket? 2014-02-12 09:12:07 ups, it's both in main and testing!? 2014-02-12 09:16:05 i think it should be removed from testing 2014-02-12 09:16:10 i think i moved it already 2014-02-12 21:22:22 ncopa, ipsec crashes on start 2014-02-12 21:22:26 with libreswan 2014-02-12 21:22:39 Feb 12 06:12:27 basalt authpriv.warn pluto[5700]: nss directory plutomain: /etc/ipsec.d 2014-02-12 21:22:39 Feb 12 06:12:27 basalt authpriv.warn pluto[5700]: NSS readonly initialization failed (err -8015) 2014-02-12 21:22:58 then if you try to stop it with rd-service, you get 2014-02-12 21:22:59 Feb 12 06:12:33 basalt daemon.err /etc/init.d/ipsec[5715]: start-stop-daemon: fopen `': No such file or directory 2014-02-12 21:32:41 first problem solved by doing "certutil -N -d /etc/ipsec.d" 2014-02-12 21:32:44 then restarting 2014-02-12 21:32:48 stop problem still there 2014-02-13 02:14:59 (1/1) Upgrading dovecot (2.2.10-r0 -> 2.2.11-r0) 2014-02-13 02:14:59 ERROR: Failed to extract etc/dovecot/dovecot.conf: Permission denied 2014-02-13 02:14:59 wget: short write 2014-02-13 02:14:59 ERROR: dovecot-2.2.11-r0: Permission denied 2014-02-13 02:15:00 huh 2014-02-13 02:17:05 got it 2014-02-13 03:03:23 could someone please upgrade libreswan to 3.8 when they get a chance 2014-02-13 09:09:33 Mp5shooter: have you tried strongswan? 2014-02-13 09:11:47 openswan is dead upstream? 2014-02-13 09:11:55 we can just remove that aport, cant we? 2014-02-13 09:13:08 i believe openswan is forked and effectively replaced by libreswan 2014-02-13 09:13:20 so probably better delete openswan, and do libreswan 2014-02-14 00:53:07 <@ncopa> Mp5shooter: have you tried strongswan? 2014-02-14 00:53:07 no 2014-02-14 09:29:28 latest musl build stats with the new lua repo build scripts: 2014-02-14 09:29:30 main testing 2014-02-14 09:29:30 musl-x86: 1672/1731 821/1104 2014-02-14 09:29:30 musl-x86_64: 1670/1727 834/1099 2014-02-14 09:29:30 musl-armhf: 1609/1720 812/1090 2014-02-14 10:30:57 those are the packages that currently is failing with musl: http://sprunge.us/LXhb 2014-02-14 13:41:24 ACTION is going to grab a buildserver from stock and install musl-x86_64/alpine-sdk 2014-02-14 22:12:35 ok ncopa i'm trying strongswan 2014-02-15 03:17:17 kern.info kernel: [177268.836878] pki[5938]: segfault at bc1a9d8 ip 00006ee4f057a17e sp 00007e47256c8420 error 4 in libuClibc-0.9.33.2.so[6ee4f0540000+75000] 2014-02-15 03:17:20 hm 2014-02-15 03:17:22 wonder what that means 2014-02-17 05:38:30 if anyone has time could they upgrade strongswan from 5.1.1 to 5.1.2rc1? it's supposed to fix a segfault bug 2014-02-17 06:20:59 @Mp5shooter - I've submitted a patch to bump strongswan to 5.1.2rc1 (it gave me a chance to test my new alpine build environment in LXC) 2014-02-17 06:21:15 nice, thanks :) 2014-02-17 06:22:03 ;-) np 2014-02-17 06:23:54 I really wanted libreswan to work out but it looks like strongswan will be about as nice as libreswan :P 2014-02-17 09:54:17 ncopa: i see this in dmesg at boot on 2.7: http://sprunge.us/TeeR 2014-02-17 09:58:52 i suppose its apk that tries to config the tty in the chroot 2014-02-17 09:59:17 but is not allowed by grsecurity 2014-02-17 11:13:04 ncopa, I think I would update jqMobile skin to v1.4x, tdtrask has updated ACF with most features I requested 2014-02-17 11:13:27 any chance to get the 'test server' ? 2014-02-17 12:03:08 sure, but not today, sorry 2014-02-17 12:09:51 could you ping me ? I would be checking logs, if not online 2014-02-17 12:10:44 would it possible to test the ACF on it ? 2014-02-17 12:11:05 vkrishn: what do you need? 2014-02-17 12:11:34 hmmm... clandmeter I mean the test server 2014-02-17 12:12:10 yes you said test server 2014-02-17 12:12:11 when ready, can I test updated ACF there ? 2014-02-17 12:12:24 what *kind* of test server? 2014-02-17 12:13:06 running to AL infra 2014-02-17 12:13:43 to? 2014-02-17 12:14:30 on* 2014-02-17 12:15:07 you want it for testing? 2014-02-17 12:16:26 clandmeter, ncopa said ^^^ not today, so when ready ping me pls 2014-02-17 12:16:54 clandmeter, some discussion on it http://www.irclogger.com/.alpine-devel/2014-01-09#1389264922 2014-02-17 12:17:23 vkrishn: I understand 2014-02-17 12:17:46 i would like to know what kind of testing 2014-02-17 12:18:53 im setting someting up 2014-02-17 12:19:00 would install ACF from edge on it, add skin jqMoble v1.4, test via URL then forward the skin to tdtrask :-) 2014-02-17 12:19:44 to be added in ACF 2014-02-17 12:19:50 ok, im building the infra now. just looking at how to process this correctly. 2014-02-17 12:20:12 :)) 2014-02-17 12:59:00 mp5shooter: do you have the exact commit that fixes the segfault? 2014-02-17 12:59:16 i'd rather just add that patch 2014-02-17 13:02:00 ncopa: can we/you cherry pick alpine-keys to 2.7? 2014-02-17 13:04:57 for the musl builders keys? 2014-02-17 13:05:00 hm 2014-02-17 13:05:08 yes 2014-02-17 13:05:13 we need add those to lxc-alpine too then 2014-02-17 13:05:16 try to isntall lxc musl 2014-02-17 13:06:09 lxc takes your current keys doesnt it? 2014-02-17 13:06:22 or it falls back 2014-02-17 13:06:40 vkrishn: too bad I arrived here after you already left 2014-02-17 13:06:49 but, you said you review the backlog ... 2014-02-17 13:07:44 ACTION created acf-jquery-mobile git repo last week 2014-02-17 13:08:08 and I have a bunch of updates to ACF in the git repos that have not yet made it into edge packages 2014-02-17 13:08:20 please start working with the git HEAD 2014-02-17 13:08:41 ncopa: does that cherry pick *fix checksum* count too? :) 2014-02-17 13:08:46 I will try to coordinate with you, because I have been slowly working on the jqmobile skin 2014-02-17 13:09:07 clandmeter: hm... :) 2014-02-17 13:09:31 i was actually thinking of merge it but the would the 'cherry-picked from ...' message be wrong 2014-02-17 13:11:26 ACTION wonders why lxc cannot use lvm by default 2014-02-17 13:28:48 ncopa: will you wait for lxc stable before edge update? 2014-02-17 13:29:18 i am trying to fix lxc things upstream first 2014-02-17 13:29:31 then pick things 2014-02-17 13:29:42 current lxc-create depends on udev 2014-02-17 13:29:49 if you use lvm 2014-02-17 13:29:50 oh? 2014-02-17 13:30:06 udevadm setlle 2014-02-17 13:30:07 ok... i havent really used lvm with lxc 2014-02-17 13:30:39 it should not be triggered if udev is not installed/running though 2014-02-17 13:31:09 but i see 1.0 converted everything to c instead of bash/sh 2014-02-17 13:31:47 i could accept if a nice and clean workaround patch 2014-02-17 13:32:26 replace it with sleep 1 :D 2014-02-17 13:32:39 always? 2014-02-17 13:33:08 if [ -x /sbin/udevadm ]; then udevadm settl; fie 2014-02-17 13:33:25 ... else mdev -s; fi 2014-02-17 13:33:27 or so 2014-02-17 13:33:53 ok 2014-02-17 13:33:56 i think i can do that 2014-02-17 13:49:21 ncopa: are you ok for added util-linux to its deps? its needed for mkfs 2014-02-17 13:50:00 and lvm2 is needed too i suppose... 2014-02-17 13:50:14 would be be possible to have a supbackage lxc-lvm or so 2014-02-17 13:50:19 which pulls in those deps? 2014-02-17 13:50:22 or 2014-02-17 13:50:31 installif? 2014-02-17 13:50:41 yeah, i was thinking install_if too 2014-02-17 13:51:02 install_if is slightly tricky though 2014-02-17 13:51:43 i'd still like to have an lxc-lvm subpackage 2014-02-17 13:51:57 with the deps collected 2014-02-17 13:52:18 ok, so just lvm and util-linux i guess 2014-02-17 13:52:28 yeah 2014-02-17 13:52:31 sounds good 2014-02-17 13:53:20 then we could consider to add: install_if="lxc=$pkgver-r$pkgrel lvm2" 2014-02-17 13:53:28 so if you apk add lvm2 lxc 2014-02-17 13:53:40 it'll pull in lxc-lvm (and coreutils...) 2014-02-17 13:53:53 and util-linux i mean 2014-02-17 13:54:15 first step is still add lxc-lvm subpackage 2014-02-17 13:54:27 ill do that and patch lxc-create 2014-02-17 13:54:33 good 2014-02-17 13:54:34 thanks! 2014-02-17 14:09:51 algitbot: wakeup 2014-02-17 14:17:40 hmm 2014-02-17 14:18:06 ncopa: abuild doesnt check depends existence? 2014-02-17 14:18:33 i think it does? 2014-02-17 14:18:35 in subpkg depends=lvm does not fail 2014-02-17 14:18:54 right, you need to add lvm2 to makedepends in addition 2014-02-17 14:19:21 abuild has no chance to know what you set in subpkg split function til you actually run it 2014-02-17 14:20:09 i dont get it. when abuild is finished it has run the split function? 2014-02-17 14:21:07 first it compiels everthying, the it runs package() 2014-02-17 14:21:15 then it runs split functions (for subpkgs) 2014-02-17 14:21:40 yes, but it doesnt fail at all 2014-02-17 14:21:43 only on install of pkg 2014-02-17 14:22:37 it is because it does not install any aditional packages (the depends in subpkg) after the package is split 2014-02-17 14:22:48 why would it? 2014-02-17 14:22:58 im not saying it should 2014-02-17 14:23:05 but it doesnt check either 2014-02-17 14:23:10 correct 2014-02-17 14:23:25 becase if it checks it has to be after it has happened 2014-02-17 14:23:27 hm 2014-02-17 14:23:32 i got used of abuild correcting me all of the time :) 2014-02-17 14:23:43 i suppose we could run a postcheck but.. 2014-02-17 14:24:06 you might also have introduced a bootstrap problem 2014-02-17 14:24:17 in case lxc is built before util-linux when bootstrapping 2014-02-17 14:24:41 abuild has no way of knowing that util-linux and lvm2 should be built before lxc 2014-02-17 14:25:02 so to help abuild, you should add util-linux and lvm2 to makedepends too 2014-02-17 14:25:22 i suppose abuild should check that 2014-02-17 14:25:24 in postcheck 2014-02-17 14:25:55 ok 2014-02-17 14:29:33 ncopa: http://wiki.strongswan.org/projects/strongswan/repository/revisions/079e6c2b04faa8cd09478662a255cc87acfaede1 2014-02-17 14:33:08 related issue says debian but they think I'm having the same problem on alpine 2014-02-17 14:33:17 very likely yes 2014-02-17 14:33:27 i am cherrypicking that commit 2014-02-17 14:33:41 ok 2014-02-17 14:35:27 Mp5shooter: can you test if it fixes your isse from edge repo? 2014-02-17 14:35:31 sure 2014-02-17 14:35:32 then I'll add it to v2.7 too 2014-02-17 14:35:39 thanks! 2014-02-17 14:35:43 thank you :P 2014-02-17 14:38:47 bah 2014-02-17 14:38:49 segfault 2014-02-17 14:39:13 i'll try some stuff when i get home 2014-02-17 14:39:46 and apk version says its strongswan-5.1.1-r2? 2014-02-17 14:40:11 yep 2014-02-17 14:40:13 just upgraded 2014-02-17 14:40:18 kernel: [391029.213923] pki[3805]: segfault at ffffffffdec2fe88 ip 000077df8276717e sp 00007f6da4138600 error 5 in libuClibc-0.9.33.2.so[77df8272d000+75000] 2014-02-17 14:40:20 ACTION shrugs 2014-02-17 14:40:29 hm 2014-02-17 14:40:51 i suppose the prper way to find out is build iwth debugging symbols 2014-02-17 14:40:59 and generate a core dump 2014-02-17 14:52:27 time for work, see you later 2014-02-17 15:12:51 vkrishn: see http://git.alpinelinux.org/cgit/tdtrask/acf-skins-jqmobile/ for first cut based on your previous work 2014-02-17 19:26:13 Does the build server keep the last few linux-headers installed ? (e.g 3.10 & 3.12) 2014-02-17 23:40:44 can anyone help me building as a root user? 2014-02-18 06:26:26 where does ABUILD install the dependencies during the build process ? (I'm building a package with a kernel module & need the location for ./configure) 2014-02-18 07:06:11 ncopa: I'm still having that weird-ass apk upgrade issue I mentioned in the mailing list 2014-02-18 07:12:43 Xe: hi 2014-02-18 07:12:54 please remind me what apk upgrade issue? 2014-02-18 07:13:13 genesect confangle (git:master) O > sudo apk upgrade 2014-02-18 07:13:15 ERROR: unsatisfiable constraints: 2014-02-18 07:13:17 Huh? Error reporter did not find the broken constraints. 2014-02-18 07:13:49 I believe its 2 bugs 2014-02-18 07:14:01 1) a bug in apk-tools (at least in error reposrting) 2014-02-18 07:14:16 any progress on them yet? I haven't upgraded this server in 2 months 2014-02-18 07:14:25 2) some indirect circular dependency in aports 2014-02-18 07:14:40 Xe: do you have audacious installed? 2014-02-18 07:14:44 ncopa: Nope 2014-02-18 07:15:04 can you please paste the /etc/apk/world and I'll try to figure out what causes it 2014-02-18 07:15:06 http://git.io/WAORNg 2014-02-18 07:16:23 weechat-nightly codemarkov-nightly 2014-02-18 07:16:27 are those homebuilt? 2014-02-18 07:16:56 yeah 2014-02-18 07:17:11 latter has no dependencies 2014-02-18 07:17:19 former is weechat built from git 2014-02-18 07:17:25 got it 2014-02-18 07:17:40 and nodejs was recently moved to "unmaintained" 2014-02-18 07:18:15 fabled: do we have some way to list all installed packages that are not available from any repo? 2014-02-18 07:19:16 maybe apk policy 2014-02-18 07:19:39 Xe: can you please try: apk policy $(cat /etc/apk/world) 2014-02-18 07:20:03 http://git.io/2PrygA 2014-02-18 07:25:01 hm 2014-02-18 07:25:20 you dont have testing repo anymore? 2014-02-18 07:25:30 in etc/apk/repositories 2014-02-18 07:26:03 http://repos.lax-noc.com/alpine/edge/main 2014-02-18 07:26:05 #http://repos.lax-noc.com/alpine/edge/testing 2014-02-18 07:26:07 #http://equine.vacantminded.com/repo/testing/ 2014-02-18 07:26:09 /shell -o cat /etc/apk/repositories 2014-02-18 07:26:28 ok 2014-02-18 07:26:30 lets do this 2014-02-18 07:26:44 this is gonna be good 2014-02-18 07:26:48 take every package in world and apk add -U it 2014-02-18 07:26:55 so like 2014-02-18 07:27:00 IFS=' 2014-02-18 07:27:02 ' 2014-02-18 07:27:18 for package in $(cat /etc/apk/world); do 2014-02-18 07:27:28 for i in $(cat /etc/apk/world); do apk add -U $i || failed="$failed" 2014-02-18 07:27:29 sudo apk add -U $package 2014-02-18 07:27:31 done 2014-02-18 07:27:35 for i in $(cat /etc/apk/world); do apk add -U $i || failed="$failed $i"; done 2014-02-18 07:27:42 echo $failed 2014-02-18 07:27:50 we want get a list of those who failed 2014-02-18 07:27:57 ad back the testing first 2014-02-18 07:28:12 then we'll end up with a shorter list 2014-02-18 07:28:28 ncopa: it's failing on every one 2014-02-18 07:28:37 every? h,, 2014-02-18 07:28:46 even like alpine-base? 2014-02-18 07:29:07 bah 2014-02-18 07:29:18 apk add -u 2014-02-18 07:29:20 not -U 2014-02-18 07:29:49 yeah 2014-02-18 07:29:52 still failing 2014-02-18 07:29:55 alpine-base 2014-02-18 07:29:57 [sudo] password for xena: 2014-02-18 07:29:59 fetch http://repos.lax-noc.com/alpine/edge/main/x86_64/APKINDEX.tar.gz 2014-02-18 07:30:01 ERROR: unsatisfiable constraints: 2014-02-18 07:30:03 Huh? Error reporter did not find the broken constraints. 2014-02-18 07:31:37 was that with -u and not -U? 2014-02-18 07:32:02 do you have testing repo in there too? 2014-02-18 07:32:56 I have -u and testing 2014-02-18 07:33:02 still failing every package 2014-02-18 07:33:17 ok 2014-02-18 07:33:27 and apk version -v 2014-02-18 07:33:54 http://git.io/DtkSTA 2014-02-18 07:35:46 its mostly likely caused by something that has one of those ? 2014-02-18 07:36:08 hm 2014-02-18 07:36:17 can you apk del abuildhelper? 2014-02-18 07:36:54 nope 2014-02-18 07:37:03 same error? 2014-02-18 07:37:18 yep 2014-02-18 07:38:06 the only time i have actually seen this error myself is when there is an unresolved intall_if 2014-02-18 07:38:44 when package X gets pulled in due to it has install_if 2014-02-18 07:39:02 but then it cannot resolve package X's dependencies 2014-02-18 07:39:10 hmm 2014-02-18 07:39:10 because some of the dependencies is missing 2014-02-18 07:39:36 as it is 2014-02-18 07:39:46 I can't reliably add or del packages 2014-02-18 07:40:01 I mean 2014-02-18 07:40:07 Hmm 2014-02-18 07:40:15 does /etc/apk/world allow comments? 2014-02-18 07:40:24 i think not 2014-02-18 07:40:29 darn 2014-02-18 07:40:35 but you can cp /etc/apk/world /etc/apk/world.backup 2014-02-18 07:40:42 if it did I could comment out all the custom packages lol 2014-02-18 07:41:13 ncopa: feature request? 2014-02-18 07:41:15 all the packages that apk version shows ? 2014-02-18 07:41:17 yeah 2014-02-18 07:41:18 yeah 2014-02-18 07:41:48 fabled: is it difficult to add support comments in /etc/apk/world? 2014-02-18 07:41:56 eg, skip lines that starts with # 2014-02-18 07:42:18 https://gist.github.com/lyska/630a58cd1782fa5650d5#file--L222 2014-02-18 07:42:25 this line intrigues me in a way 2014-02-18 07:42:37 I remember I did build some python version 2014-02-18 07:42:49 but I don't think I built python sqlite as its own package 2014-02-18 07:43:21 make a backup of world file 2014-02-18 07:43:23 and try edit it 2014-02-18 07:43:42 okay 2014-02-18 07:44:20 you could try one by one package 2014-02-18 07:44:26 remove it 2014-02-18 07:44:29 and try apk upgrade 2014-02-18 07:44:39 that way you might find what triggers it 2014-02-18 07:45:38 for i in $(cat /etc/apk/world.backup); do sed "/^$i/d" /etc/apk/world.backup > /etc/apk/world; .... 2014-02-18 07:45:40 or similar 2014-02-18 07:46:18 does that add every line to /etc/apk/world one by one? 2014-02-18 08:04:32 sed line remove one line 2014-02-18 08:05:00 sed '/searchpattern/d' 2014-02-18 08:05:13 deletes the lines which matches searchpattern 2014-02-18 12:47:20 how can I prevent a pkg (which requires libtool to build) - from pulling in dependencies for binutils / binutils-libs / libgomp / pkgconf / pkgconfig / mpfr3 / mpc1 / gcc ??? 2014-02-18 12:52:56 *.a & *.la are already being deleted from /usr/lib 2014-02-18 13:01:43 ok - rm autoconf / automake from makedepends 2014-02-18 13:04:25 the only question I have is where in the build environment are the dependencies held ? (or can APKBUILD use /usr/src/modules.....) ? 2014-02-19 05:55:31 does anyone have a grsecurity.patch for the 3.10.28 kernel ? (grsecurity.net has 3.13 & is too new - any 3.10.x patch will do) 2014-02-19 05:56:25 BitL0G1c, ncopa has been doing unofficial rebases 2014-02-19 05:57:31 BitL0G1c, http://git.alpinelinux.org/cgit/aports/plain/main/linux-grsec/grsecurity-2.9.1-3.10.28-unofficial.patch?h=2.7-stable 2014-02-19 05:57:53 great - many thx 2014-02-19 05:58:20 BitL0G1c, ncopa's rebase git tree is at https://github.com/ncopa/linux-stable-grsec 2014-02-19 05:58:48 thank you 2014-02-19 09:23:33 I've built a pkg with a separate kernel module pkg (I RTFM & followed the structure for "open-vm-tools / open-vm-tools-grsec) - but running "abuild -r" on the module APKBUILD doesn't generate a build - is this by design ? 2014-02-19 09:35:16 musl-x86 repos now have openjdk6 2014-02-19 09:35:26 BitL0G1c: i dont think its by design 2014-02-19 09:35:30 sounds like some kind of bug 2014-02-19 09:35:46 do you have arch= set? 2014-02-19 09:37:18 arch is blank 2014-02-19 09:37:28 "" 2014-02-19 09:38:08 shouldn't it be set? 2014-02-19 09:39:30 that's better - it should probably be set in "open-vm-tools-grsec" too 2014-02-19 09:39:54 I used that APKBUILD as the template 2014-02-19 09:49:19 setting it blank is a way to disable the package 2014-02-19 09:49:35 i probably disabled open-vm-tools-grsec due to it broke with 3.12 kernel or so 2014-02-19 09:49:44 and i wanted push the kernel 2014-02-19 09:52:07 ah right 2014-02-19 09:53:12 #2687 2014-02-19 09:53:35 i think enabling acl in coreutils broke things 2014-02-19 09:53:49 does anyone need it? 2014-02-19 09:54:09 looks like it was fabian_a that enabled it 2014-02-19 10:08:58 yup 2014-02-19 10:09:19 coreutils' install FILE /dir 2014-02-19 10:09:34 will not install it as 'root' 2014-02-19 10:09:55 while busybox's install does 2014-02-19 10:13:00 one last question - I should probably not be passing "/usr/src/linux-headers-${_abi_release}" to ./configure - how do I point to the source inside the build environment ? (linux-grsec-dev is a makedepends) 2014-02-19 10:16:51 depends on the package you build 2014-02-19 10:16:56 they all do things differently 2014-02-19 10:17:51 it works using /usr/src.... - I just thought this is actually pointing to my machine & not the build environment 2014-02-19 10:20:09 i think you can point source dir to /usr/src/linux-headers-$_abi_release 2014-02-19 10:20:39 ok 8-) 2014-02-19 10:21:03 yeah 2014-02-19 10:21:22 thats what we do in xtables-addons-grsec and ipfw-grsec 2014-02-19 10:24:09 ok - I've built openvswitch2 & openvswitch2-grsec - will submit a patch a little later 2014-02-19 10:26:47 does it require 3rd party kernel module? 2014-02-19 10:27:01 I could never get "openvswitch" to work - segfaullts - think it was built against the 3.12 kernel - openvswitch currently only build against max 3.10 2014-02-19 10:27:03 or, *why* does it require 3rd party 2014-02-19 10:27:36 it's their own module - performance will be better with the module 2014-02-19 10:28:02 i wonder why its not included in mainline kernel 2014-02-19 10:28:08 you can build just the userland - or userland+ module 2014-02-19 10:30:45 maybe because it's still subject to change - the confd changed a little 2014-02-19 10:31:20 in version 2.0.0+ 2014-02-19 10:34:10 "as of linux 3.3 it is part of the mainline kernel" 2014-02-19 10:35:13 http://git.openvswitch.org/cgi-bin/gitweb.cgi?p=openvswitch;a=blob_plain;f=FAQ;hb=HEAD 2014-02-19 10:39:23 we have an openvswitch package already 2014-02-19 10:40:02 maybe should just upgrade it to 2.0? 2014-02-19 10:40:29 I can do that instead 2014-02-19 10:40:51 that would be better yes 2014-02-19 10:41:14 ok 2014-02-19 12:19:38 ncopa: what's left on missing packages list? 2014-02-19 12:24:13 barthalion: http://sprunge.us/jjSj 2014-02-19 12:24:23 I think I'll have a stab at qemu next 2014-02-19 12:26:02 I'll take a look at varnish 2014-02-19 12:40:08 ncopa: any reason why our musl doesn't provide sys/sysctl.h? 2014-02-19 12:40:27 dunno? maybe its not in posix? 2014-02-19 12:40:38 musl has it, it's just not here 2014-02-19 12:40:54 % pacman -Ql musl | grep sysctl 2014-02-19 12:40:54 musl /usr/musl/include/sys/sysctl.h 2014-02-19 12:42:24 seems it doesn't do much though… 2014-02-19 12:42:38 int sysctl (int *, int, void *, size_t *, void *, size_t) 2014-02-19 12:43:18 I'll handle it with some ifdefs 2014-02-19 13:32:01 is there a way to blacklist incorrect package dependencies in APKBULD ? 2014-02-19 13:36:54 depends="!badpkg" 2014-02-19 13:36:57 great !!!! 2014-02-19 13:36:57 will make package install fail if 'badpkg' is installed 2014-02-19 13:37:12 you can add it to makedepends if you want specific package not be installed when you build package 2014-02-19 13:39:54 ok - will try that first (binutils / pkgconf / gcc / mpfr3 / mpc1) - to blacklist 2014-02-19 13:40:24 !? 2014-02-19 13:40:28 that sounds wrong... 2014-02-19 13:40:52 it will not build without binutils 2014-02-19 13:41:02 what is the exact problem you are trying to solve? 2014-02-19 13:41:15 installing openvswitch also installs those 2014-02-19 13:42:05 on 2.0.1 it doesn't - but on the latest GIT (which uses libtool) - they no get installed with openvswitch 2014-02-19 13:42:19 *they now 2014-02-19 14:31:49 aha 2014-02-19 14:32:05 strongswan dev found out why ipsec pki, part of strongswan is segfaulting 2014-02-19 14:32:09 on alpine 2014-02-19 14:32:24 <+ecdsa> Mp5shooter: I was able to reproduce the crash. It is caused because uclibc doesn't implement fmemopen(), which is used by the sshkey plugin. Apparently, strndup is also missing. A workaround for the pki crash is to configure with --disable-sshkey, but the strndup issue is not fixed by that (I will look at this now). 2014-02-19 14:33:10 nice! 2014-02-19 14:33:29 strndup should not be missing 2014-02-19 15:45:20 ncopa, anything on websocket/mqtt ? 2014-02-19 15:45:57 found some useful links at 0mq ML, https://dpaste.de/XX6y 2014-02-19 15:48:00 tdtrask, thanks for 'acf-skins-jqmobile', waiting for test server to get onto it 2014-02-19 15:48:11 vkrishn: hi 2014-02-19 15:48:20 hi 2014-02-19 15:48:29 which ports do you need for access? 2014-02-19 15:48:43 and which repo? 2014-02-19 15:49:22 global http: for testing, acesss could be restricted to my ip 2014-02-19 15:49:34 all test on 'edge' 2014-02-19 15:50:49 ssh access too 2014-02-19 15:50:59 no telnet? 2014-02-19 15:51:15 :p 2014-02-19 15:51:24 :) 2014-02-19 15:51:27 clandmeter: can I have rlogin access? :3 2014-02-19 15:54:11 'vkrishn-edge-x86_64' created 2014-02-19 15:56:26 need email to send access details ? 2014-02-19 15:57:16 vkrishn: pm me your ssh key 2014-02-19 15:59:50 vkrishn: you have one? 2014-02-19 16:00:04 yes, use it on github 2014-02-19 16:00:30 did you get it, I set a file ? 2014-02-19 16:00:35 can you pm me your pub key? 2014-02-19 16:00:49 or email 2014-02-19 17:42:17 clandmeter, http access to 'vkrishn-edge-x86_64' allowed ? 2014-02-19 22:37:57 vkrishn: you can use ssh socks with browser or wait for me tomorrow. 2014-02-20 00:01:37 hey ncopa, strongswan devs pushed a commit to their git which should fix the segfaults 2014-02-20 07:34:59 clandmeter, now all set and using ssh socks, thanks 2014-02-20 07:53:53 ok 2014-02-20 07:53:56 good news: 2014-02-20 07:54:04 openjdk6 is built and pushed to musl-x86 and musl-x86_64 2014-02-20 07:54:14 openjdk7 is also built, and i'm doing some testing 2014-02-20 07:54:27 turns out the fix for musl, might make openjdk7 also work with uclibc 2014-02-20 08:00:58 fabled, do you have any note about how to set-up a musl env? 2014-02-20 08:01:34 i can install a vm with musl iso and build a build-env 2014-02-20 08:03:43 i have iso images that you can use at http://dev.alpinelinux.org/~tteras/alpine-140103-x86.iso and http://dev.alpinelinux.org/~tteras/alpine-140103-x86_64.iso 2014-02-20 08:03:47 you can use those 2014-02-20 08:03:54 though - mkinitfs might be still broken a bit 2014-02-20 08:04:04 chroot / lxc works nicely 2014-02-20 08:04:30 brb 2014-02-20 08:05:10 lxc is fantastic - I use it as my development environment 2014-02-20 08:05:29 fabled: thx 2014-02-20 08:16:36 lxc is indeed nice 2014-02-20 08:17:21 I was using KVM to build in - not any more 2014-02-20 08:17:50 i dont like my current setup 2014-02-20 08:18:19 when using lvm, would prefer it stores container config to etc 2014-02-20 08:19:09 rather than /var/lib/guest/config 2014-02-20 08:19:19 correct 2014-02-20 08:19:33 use datadir only for storage if needed 2014-02-20 08:20:04 now if i lbu_include /var/lib/lxc 2014-02-20 08:20:04 i think you can change that behavior in /etc/lxc/lxc.conf 2014-02-20 08:20:22 and i forget to use lvm.... 2014-02-20 08:20:23 :D 2014-02-20 08:21:03 i was looking into that webinterface on github for lxc 2014-02-20 08:21:21 for small personal setup, that could be nice to have. 2014-02-20 08:21:49 but i think we are missing some python modules 2014-02-20 08:23:37 py-libvirt exists in alpine 2014-02-20 08:28:51 i never used lvm with lxc 2014-02-20 08:29:05 i kinda like that the containers share filesystem 2014-02-20 08:29:17 but thats for my dev envs 2014-02-20 08:29:54 seems nice, http://sheepdog.github.io/sheepdog/ 2014-02-20 08:31:33 the only advantage of installing an LXC guest into an LVM volume is for faster machine cloning - you could get the same benefit with /var on a btrfs partition - lxc will automatically use a snapshot to clone 2014-02-20 08:31:59 interesting 2014-02-20 08:37:33 openjdk7 built for uclibc too 2014-02-20 08:46:15 \o/ 2014-02-20 08:46:23 nice job fabled 2014-02-20 08:46:50 ncopa, i'm wondering if i should just go ahead and push openjdk7 to main (or testing?) 2014-02-20 08:47:00 self-build succeeds so it looks quite stable 2014-02-20 08:47:17 add it to testing 2014-02-20 08:47:23 oh 2014-02-20 08:47:24 hm 2014-02-20 08:47:29 we save compile job 2014-02-20 08:47:34 if we move it directly to main... 2014-02-20 08:47:34 yeah. it's a beast 2014-02-20 08:47:40 takes a while to build 2014-02-20 08:47:51 i tested it on edge-x86_64 and edge-musl-x86 2014-02-20 08:47:52 both work 2014-02-20 08:48:00 helloworld.java works? 2014-02-20 08:48:07 yes, but also it can self-compile. 2014-02-20 08:48:26 as in, bootstrapped openjdk is used to rebuild self, like gcc bootstrap 2014-02-20 08:48:41 i'm ok to push it directly to main to save the extra compiling 2014-02-20 08:48:47 will do 2014-02-20 08:50:24 dhcpcd fails 2014-02-20 08:50:42 due to the musl patches? 2014-02-20 08:50:44 I think most dependency for sheepdog, like corosync...etc are in aports 2014-02-20 08:51:21 if anything like sheepdog is not in AL, I could set a request in bug.a.o 2014-02-20 08:51:25 hm... i tested only on musl... 2014-02-20 08:51:51 fabled: what did we say about build it local first! 2014-02-20 08:51:55 :p 2014-02-20 08:52:18 :-( 2014-02-20 08:52:27 i did built it local but only musl 2014-02-20 08:52:42 i suppose we should swap edge to musl soonish 2014-02-20 09:04:24 i'll fix it 2014-02-20 10:13:25 all of a sudden i get various: dlsym(acl_get_fd): Symbol not found: acl_get_fd 2014-02-20 10:13:25 dlsym(acl_get_file): Symbol not found: acl_get_file 2014-02-20 10:13:25 dlsym(acl_set_fd): Symbol not found: acl_set_fd 2014-02-20 10:13:25 dlsym(acl_set_file): Symbol not found: acl_set_file 2014-02-20 10:13:25 dlsym(acl_get_fd): Symbol not found: acl_get_fd 2014-02-20 10:13:30 from fakeroot 2014-02-20 10:13:33 i wonder what happened 2014-02-20 10:14:36 ah 2014-02-20 10:14:37 i know 2014-02-20 10:14:46 (2/101) Downgrading fakeroot (1.20-r1 -> 1.20-r0) 2014-02-20 12:14:12 ncopa: can i trash my uclibc builder and go musl? 2014-02-20 12:14:54 i'd guess so 2014-02-20 12:16:21 i think ill more some stuff to my new box 2014-02-20 12:16:32 has a few more cpu's 2014-02-20 12:16:57 if it runs stable, we can think about moving other infra 2014-02-20 12:17:35 i still need to go over the other servers and see if it has anything usefull. 2014-02-20 12:54:57 monit now at v5.7, update pls 2014-02-20 13:22:15 thanks 2014-02-21 02:19:35 ncopa: around? 2014-02-21 09:50:46 anyone using Btrfs on alpine? 2014-02-21 12:37:14 clandmeter, hi 2014-02-21 12:38:44 hi 2014-02-21 12:38:53 clandmeter, would be ok if you plan to upgrade to lxc 1.0x for "vkrishn-edge-x86_64", I have backups 2014-02-21 12:39:56 vkrishn: i made a backup of your guest 2014-02-21 12:40:05 if you are intested 2014-02-21 12:40:06 thanks :) 2014-02-21 12:40:25 im trying to get it back online 2014-02-21 12:40:50 some trouble in btrfs and also alpine-iso 2014-02-21 14:18:56 vkrishn: it should be back online 2014-02-21 14:19:40 ok 2014-02-21 14:20:15 please try if you can login 2014-02-21 14:20:21 i put the backup back in /root 2014-02-21 14:32:17 log in ok, thanks 2014-02-22 21:10:41 ncopa: around? 2014-02-23 12:14:11 tdtrask, around ? 2014-02-23 14:50:59 ncopa: around? 2014-02-23 16:39:05 vkrishn: no, I wasn't around :( 2014-02-23 16:41:27 ACTION will be mostly available during business hours (EST) tomorrow 2014-02-24 01:11:59 ncopa: around? 2014-02-24 07:17:44 Mp5shooter: hi 2014-02-24 07:17:48 im here 2014-02-24 07:17:52 hey ncopa 2014-02-24 07:17:53 :P 2014-02-24 07:17:54 was sick this weekend sorry 2014-02-24 07:18:02 oh haha 2014-02-24 07:18:03 no problem 2014-02-24 07:18:12 http://wiki.strongswan.org/issues/516 2014-02-24 07:18:20 they fixed the pki stuff 2014-02-24 07:19:41 nice! 2014-02-24 07:19:47 oh this is really nice 2014-02-24 07:19:50 yeah :P 2014-02-24 07:20:12 i bet this is good for musl too 2014-02-24 07:20:27 oh nice 2014-02-24 07:20:43 i havent dig into the sources but it looks like defining _GNU_SOURCE is the right thing to do 2014-02-24 07:21:05 :o 2014-02-24 07:23:49 is the fix included in the rc2? 2014-02-24 07:24:19 no 2014-02-24 07:24:22 nope 2014-02-24 07:24:25 looks like not 2014-02-24 07:24:32 I think rc2 was a couple of days before the fix 2014-02-24 07:25:01 i suppose 5.1.2 is around the corner 2014-02-24 07:25:43 probably 2014-02-24 07:29:15 time for bed if i want to wake up on time :P good night 2014-02-24 07:29:28 good night 2014-02-24 15:47:31 build-2-5: retry 2014-02-24 15:51:58 files from v2.5.4-244-ga893bba uploaded 2014-02-24 15:58:39 files from v2.5.4-244-ga893bba uploaded 2014-02-24 15:59:33 files from v2.4.11-188-g44517e8 uploaded 2014-02-24 16:02:26 ncopa: I'll push new ffmpeg to main in a moment 2014-02-24 16:22:32 ncopa: what's left on the list? 2014-02-24 16:24:33 files from v2.4.11-189-g89d8ce6 uploaded 2014-02-25 07:00:15 morning 2014-02-25 07:00:22 its patch tuesday 2014-02-25 07:00:32 i'll try fix bugs today 2014-02-25 07:09:41 barthalion: http://sprunge.us/CjLX 2014-02-25 07:09:51 hm 2014-02-25 07:09:53 subversion? 2014-02-25 07:13:54 http://sprunge.us/APjG 2014-02-25 07:36:39 files from v2.5.4-245-ga83cec9 uploaded 2014-02-25 07:37:49 files from v2.5.4-245-ga83cec9 uploaded 2014-02-25 08:29:47 seems like 3.12 kernel is announced to be longterm 2014-02-25 08:30:05 i wonder if we should try stick to 3.12 for v2.8/v3.0 2014-02-25 08:51:38 reminds me, alpine-iso generates wrong syslinux.conf 2014-02-25 08:52:37 ncopa: i think you recently changed something with boot generation (change kernel filename) 2014-02-25 08:52:51 in edge? 2014-02-25 08:52:54 or maybe my local conf was still old 2014-02-25 08:52:58 yes edge 2014-02-25 08:53:02 i think i did change something in edge yes 2014-02-25 08:53:20 due to being able to use /boot on vfat 2014-02-25 08:53:25 for EFI 2014-02-25 08:53:25 i know i had to rename kernel in syslinux.conf 2014-02-25 08:53:35 before it was grsec 2014-02-25 08:53:36 hm 2014-02-25 08:53:50 shouldnt update-extlinux be able to handle that? 2014-02-25 08:54:12 or is it a tmpfs setup? 2014-02-25 08:54:12 i created an iso 2014-02-25 08:54:14 ok 2014-02-25 08:54:15 yes 2014-02-25 08:54:25 that will need some modification 2014-02-25 08:54:31 btw 2014-02-25 08:54:42 can we do something like a monthly edge iso? 2014-02-25 08:54:55 that sounds like a good idea yes 2014-02-25 08:55:11 mini is enough for most user cases i guess 2014-02-25 08:55:17 yes 2014-02-25 08:55:21 edge snapshots 2014-02-25 08:55:30 or we could call it _alpha 2014-02-25 08:55:39 and the version shows 2.8 2014-02-25 08:55:47 2.8_alpha or so 2014-02-25 08:55:49 btw 2014-02-25 08:56:01 well, depends if its going to be 2.8 :) 2014-02-25 08:56:02 i have been thinking on how to do the migration to musl 2014-02-25 08:56:04 yes 2014-02-25 08:56:18 we need decide what to do soonish 2014-02-25 08:56:40 if we stick to uclibc for the releas in May, and do 2.8 2014-02-25 08:56:49 ncopa: 3.12 sounds reasonable 2014-02-25 08:56:54 then we cannot switch edge to musl til after that 2014-02-25 08:57:22 if we want do 3.0 and musl in may 2014, then we should switch edge to musl asap 2014-02-25 08:57:40 do we have showstoppers? 2014-02-25 08:57:55 http://sprunge.us/APjG 2014-02-25 08:58:03 + fix mkinitfs 2014-02-25 08:58:22 i think i can fix qemu soonish 2014-02-25 08:58:36 i actually got it built but it broke uclibc... 2014-02-25 08:59:02 (so if w just jump sideways to musl and leave uclibc behind we could just use that...) 2014-02-25 08:59:18 xen i dunno 2014-02-25 08:59:34 but we could leave xen behind too, for 3.0 2014-02-25 08:59:38 and make 3.0 without xen 2014-02-25 08:59:44 i'd be okish with that 2014-02-25 08:59:58 i dont know what needs to be done with xen 2014-02-25 09:00:29 ebtables needs some refactoring 2014-02-25 09:00:33 not really trivial 2014-02-25 09:00:52 same with coova-chilli 2014-02-25 09:01:08 cups-filters fails on a bug in gcc headers or so 2014-02-25 09:01:36 not trivial but still possible to fix 2014-02-25 09:01:54 libspf2 fails due to missing resolv stuff 2014-02-25 09:01:58 not sure how to fix that 2014-02-25 09:02:20 kbd just needs work, not really difficult. just much work 2014-02-25 09:02:42 drbd uses hearch_r, which is not implemented in musl 2014-02-25 09:03:00 it needs refactoring or we need add support for those things in musl 2014-02-25 09:03:15 might be we could use hsearch (non rentrant) 2014-02-25 09:03:55 i dont think the code in drbd runs in threads/parallel so we could probably make it use the posix functions 2014-02-25 09:04:19 not that hard, but you need to know a bit C prorgamming 2014-02-25 09:04:37 reiserfs and utilvserver i dont remember 2014-02-25 09:04:49 I'll take a look at reiserfs and ufsutils 2014-02-25 09:05:14 good! 2014-02-25 09:05:15 now 2014-02-25 09:05:18 the question is 2014-02-25 09:05:27 if we move to musl and 3.0 2014-02-25 09:05:35 how do we upgrade edge builders? 2014-02-25 09:05:42 what about edge users? 2014-02-25 09:06:05 we could manually update the builders 2014-02-25 09:06:17 well, edge users should be prepared for possible breakage, that's the fun 2014-02-25 09:06:17 manually replace the toolchain 2014-02-25 09:06:32 then bump pkgrel on almost everything 2014-02-25 09:06:37 to make it rebuild things 2014-02-25 09:06:50 and hope apk-tools and so:depends fixes it 2014-02-25 09:07:08 other possibility is to rename 'edge' to something else 2014-02-25 09:07:16 to 'head', 'master' or 'rolling' 2014-02-25 09:07:34 and say that from now on, edge users should use 'rolling' branch instead of 'edge' 2014-02-25 09:08:00 and once the new name/builders are up2date, we decommision 'edge' 2014-02-25 09:08:44 the current edge usres will have to manually change 'edge' to 'rolling' (or whatever we call it) and then run apk upgrade -U -a 2014-02-25 09:09:19 apk is fast so even the biggest desktop installs should be be reinstalled within 5 mins, depending on internet speed 2014-02-25 09:09:49 i think the third option is to do v2.8 and wait with musl 2014-02-25 09:10:08 I guess it's best one 2014-02-25 09:10:28 but then edge will continue as uclibc til atleast after v2.8 is out 2014-02-25 09:10:29 renaming branch seems quite ugly 2014-02-25 09:10:43 but I'm not sure if apk-tools and so:depends can handle the rebuild 2014-02-25 09:10:51 well 2014-02-25 09:11:25 in practice, we'll need to do the rename operation on the builders (except that we keep same name) 2014-02-25 09:11:37 we will have to replace the builders with new ones 2014-02-25 09:11:50 and make it rebuild things 2014-02-25 09:12:14 i think the way to do it would be: 2014-02-25 09:12:22 1. stop current edge builders 2014-02-25 09:12:31 2. set up new musl edge builders 2014-02-25 09:12:53 3. bump pkgrel on everythin thats linked to libc (bascially everything thats not noarch or kernel) 2014-02-25 09:13:11 waith til it built 2014-02-25 09:13:31 4. swap the edge builders with the new musl one 2014-02-25 09:13:53 the so: depends should make sure things does not clash 2014-02-25 09:14:15 problem will arise for those who has home built stuff 2014-02-25 09:14:22 they will have to rebuild it too 2014-02-25 09:14:52 basically, swithc to musl means complete reinstall 2014-02-25 09:21:49 finally got my arm builder online. will try to sync it to the latest aports soon. 2014-02-25 09:21:56 it'll takes quite a while. doing gcc now. 2014-02-25 09:22:06 iirc, rebuilding gcc is 3+ hours 2014-02-25 09:25:53 if we do 3.0 in May with musl 2014-02-25 09:26:12 then we should probably announce it as Beta quality 2014-02-25 09:26:26 and we should probaly give it shorter maintenance time 2014-02-25 09:26:45 in which case it doesnt really matter that we dont use a LTS kernel... 2014-02-25 09:27:04 right. means that we likely need to extend 2.7 support time too 2014-02-25 09:27:23 which is kinda okish 2014-02-25 09:27:34 it has a good kernel, 3.10 2014-02-25 09:27:46 fabled: did you get that 3.12 is announced longterm 2014-02-25 09:27:57 no 2014-02-25 09:28:10 but good 2014-02-25 09:28:24 seems like 3.12 kernel is announced to be longterm 2014-02-25 09:28:24 i wonder if we should try stick to 3.12 for v2.8/v3.0 2014-02-25 09:28:38 the problem with 3.12 is that grsecurity has already left it behind... 2014-02-25 09:28:48 so it is kinda old already 2014-02-25 09:28:59 but same case with 3.10 2014-02-25 09:29:16 or do they have extended grsec support for 3.10 ? 2014-02-25 09:29:19 no 2014-02-25 09:29:41 just that we moved to 3.10 closer to the v2.7 release 2014-02-25 09:30:02 so it was fresher when v2.7 came out 2014-02-25 09:30:04 ah, right 2014-02-25 09:30:29 i was kinda hoping for 3.14 to be LTS 2014-02-25 09:30:41 but it's not even out yet 2014-02-25 09:30:48 3.10 and 3.12 are LTS 2014-02-25 09:30:56 3.14 might be too... who knows 2014-02-25 09:31:10 but, again, my main point is, if we do 3.0, and we cut down maintenance time 2014-02-25 09:31:27 then i dont think it matters that much if we dont hit a LTS kernel for 3.0 2014-02-25 09:31:32 right 2014-02-25 09:31:42 if we do v2.8 and uclibc, then i very much want LTS kernel 2014-02-25 09:31:52 understand 2014-02-25 09:32:59 also, it seems like grsecurity's longterms kernels depend on ubuntus kernel 2014-02-25 09:33:23 so it might be an idea to keep an eye on what they do 2014-02-25 09:33:59 new ubuntu lts will most likely use 3.13 2014-02-25 09:37:29 thats what i read too 2014-02-25 09:37:36 but they are looking 3.14 too 2014-02-25 09:38:36 this is kinda cool: http://phb-crystal-ball.org/ 2014-02-25 09:38:46 the v3.14 kernel is predicted for Saturday, 2014-03-29 2014-02-25 09:38:55 so ubuntu will likely not do 3.14 kernel 2014-02-25 09:38:57 hm 2014-02-25 09:39:23 in any case, ubuntu will *not* use 3.12 kernel 2014-02-25 09:39:38 so i suppose its ok for us to upgrade the edge kernel to 3.13 2014-02-25 09:39:39 yeah, I guess the "push everything until we can" time is over in 14.04 2014-02-25 09:40:27 when you aim for 5yrs support then you really want as new as possible 2014-02-25 09:43:24 barthalion: btw, nice work on thos musl build fixes 2014-02-25 09:43:26 really nice! 2014-02-25 09:43:29 thanks! 2014-02-25 10:35:12 np 2014-02-25 10:40:18 seems that I need to add printf.h support to musl 2014-02-25 15:00:49 i've installed alpine musl on kvm 2014-02-25 15:00:57 fabled's iso 2014-02-25 15:01:17 it works smoothly if the repositories are correctly configured 2014-02-25 15:01:26 and during the installation he can access to internet 2014-02-25 15:01:43 if not, installation goes ahead but segfault on reboot 2014-02-25 15:02:11 i'll try to setup a buildenv 2014-02-25 15:03:36 I've running buildenv in systemd container, works fine 2014-02-25 15:07:43 i was not aware that systemd could be used as container (i don't really like systemd) 2014-02-25 15:07:51 http://0pointer.de/public/systemd-man/systemd-nspawn.html 2014-02-25 15:07:56 this looks something to try 2014-02-25 15:10:22 ncopa is trying to convert me to LXC, vainly :p 2014-02-25 15:10:40 just trying systemd now :) 2014-02-25 15:11:12 I still need to automate package building with nspawn 2014-02-25 15:11:15 and chroot recreation 2014-02-25 15:28:22 btw, lxc in edge is a bit broke 2014-02-25 15:28:25 nees some polishing 2014-02-25 15:28:52 speaking of lxc, I thought its scripts are self-contained 2014-02-25 15:29:13 I had to make /etc/apk here, fill the repositories list and put keys in there 2014-02-25 15:31:34 it's not very different from running manually apk.static, doing the same thing with repos and nspawn then… 2014-02-25 15:59:44 files from v2.5.4-246-gea0840f uploaded 2014-02-25 15:59:52 files from v2.5.4-246-gea0840f uploaded 2014-02-25 16:34:18 ncopa: whats wrong with lxc? 2014-02-25 16:35:05 barthalion: you needed to do that in the guest? 2014-02-25 16:36:09 no, host 2014-02-25 16:36:22 afaik, lxc-alpine copies the repository from the host 2014-02-25 16:36:42 yeah, but I'm not running Alpine, thus the problem 2014-02-25 16:40:09 barthalion: you can specify the repo 2014-02-25 17:20:21 files from v2.4.11-194-g7f7d47b uploaded 2014-02-25 17:26:24 files from v2.4.11-194-g7f7d47b uploaded 2014-02-25 18:22:51 barthalion, r u around? 2014-02-25 18:44:16 fcolista: just a little bit 2014-02-25 18:44:23 k 2014-02-25 18:44:41 when you use systemd-nspawn, u use it only as chroot right? 2014-02-25 18:44:57 or you are able to boot alpine-musl with it? 2014-02-25 18:45:06 no, I haven't tried 2014-02-25 18:45:11 k 2014-02-25 18:45:14 it should be doable if you create some devices in /dev/ 2014-02-25 18:45:24 'cause you should use systemd 2014-02-25 18:45:31 or hack mdev 2014-02-25 18:45:52 basically tty does not appears 2014-02-25 18:46:02 even creating tty[1-6] 2014-02-25 18:46:19 well, systemd doesn't notice that creation 2014-02-25 18:46:42 How are you using s-nspawn then? 2014-02-25 18:49:29 well, I use it as build environment, not container 2014-02-25 18:49:47 systemd-nspawn -D alpine /bin/sh 2014-02-25 19:02:20 k 2014-02-25 19:02:50 it's just a chroot then. I don't see any advantage in using it in such way. 2014-02-25 19:12:18 why would I need more to check if package builds? 2014-02-25 19:13:44 and well, it's not just a chroot, it gets its own namespace, and access to /sys is read only 2014-02-26 07:15:47 this is interesting: http://www.openbsdfoundation.org/gsoc2014.html#systemd 2014-02-26 08:00:00 ^^^ sorry, working on this ^^^ 2014-02-26 10:44:02 ncopa: what is wrong with current lxc? 2014-02-26 10:49:37 hi 2014-02-26 10:49:42 the init.d script start/stop 2014-02-26 10:50:16 previously we hacked stop script to check if it was busybox init and if so, we passed over a different signal 2014-02-26 10:50:24 now we need to config the stop signal in config 2014-02-27 12:53:19 chrome syntax highlighting, didnt know it existed. 2014-02-27 12:59:12 ncopa: can we add less to git? busybox less doesnt like colors. 2014-02-27 12:59:19 or maybe have git-colors 2014-02-27 13:00:46 can't you disable colors in git? 2014-02-27 13:01:02 also i think there was some patch to busybox less to add support for colors 2014-02-27 13:01:03 i 2014-02-27 13:01:11 i'd prefer not add hard dep of less 2014-02-27 13:01:44 I like having colors in git 2014-02-27 13:01:52 so do I 2014-02-27 13:01:54 apk add less 2014-02-27 13:01:57 and problem solved 2014-02-27 13:01:57 and i always have to search how to enable it 2014-02-27 13:02:06 thast not logical 2014-02-27 13:02:17 for you maybe 2014-02-27 13:02:28 not for the avarage joe like me :) 2014-02-27 13:02:31 i am interested in the patch for bb less for color support 2014-02-27 13:03:51 what about a sub package? 2014-02-27 13:04:12 atleast somebody could find the solution in aports 2014-02-27 13:34:08 i think the solution is fix busybox less 2014-02-27 13:34:18 there was a patch sent to busybox ml not too long ago 2014-02-27 13:51:24 ncopa: looks like only build-edge-x86_64 fails on main -> testing deps? 2014-02-27 13:51:43 yes 2014-02-27 13:52:05 packages in main may not depend on packages in other repo 2014-02-27 13:52:11 the other way around is ok 2014-02-27 13:52:19 yes i know 2014-02-27 13:52:25 but why x86 doesnt? 2014-02-27 13:52:37 or am i readign it wrong? 2014-02-27 13:53:29 dont know 2014-02-27 13:53:50 it might be it didnt fail for other reason 2014-02-28 09:14:59 tdtrask, ping 2014-02-28 12:25:53 alpine is mentioned: http://www.openwall.com/lists/musl/2014/02/28/1 2014-02-28 13:02:54 ncopa: "For the ones where support is complete or significant process is being made, we should highlight their support of musl in the publicity work we do elsewhere" I like this part 2014-02-28 13:02:55 =] 2014-02-28 13:09:02 me too :) 2014-02-28 13:09:05 and actually 2014-02-28 13:09:15 i believe nobody has so many packges built with musl as alpine 2014-02-28 14:42:42 vkrishn: missed you again :( 2014-02-28 15:29:41 tdtrask, anything on implementing 'label' for form elements and adding variable 'pageinfo.skindir' ? 2014-02-28 15:32:53 gtg, leave a msg here or email 2014-02-28 16:52:49 vkrishn: it's been a crazy week for me, I'll try to take a look at it on Monday 2014-02-28 16:52:57 shouldn't be difficult, just need to do it :) 2014-02-28 16:54:19 pageinfo.skindir shouldn't be necessary since you already have pageinfo.wwwprefix and pageinfo.skin 2014-02-28 16:54:48 pageinfo.skindir = pageinfo.wwwprefix .. pageinfo.skin 2014-02-28 16:56:03 ACTION has received something similar to the 'label' request from someone else too, so definitely will add it 2014-02-28 17:06:36 eg. for skin 'jqmobile', pageinfo.skindir resolves to "/usr/share/acf/www/skins/jqmobile/" 2014-02-28 17:07:05 not www path , but filesystem path 2014-02-28 17:08:15 though I can hard code "/usr/share/acf/www/skins" in template, but would seems nice if were a variable 2014-02-28 17:08:53 thanks, for working on label 2014-02-28 17:46:03 vkrishn: where are you getting pageinfo.skindir from? It's not defined in any of my code (although it was in the original template files you provided me) 2014-02-28 17:46:18 are you saying that pageinfo.wwwprefix is wrong? 2014-02-28 17:47:36 maybe you have a mix of new and old code on the box? 2014-02-28 18:13:23 tdtrask, I was wanting to have variable 'pageinfo.skindir', if possible, presently its not there , I know 2014-02-28 18:14:27 would it be possible ? 2014-02-28 18:20:17 I had defined, skindirurl = pageinfo.wwwprefix..pageinfo.skin 2014-02-28 18:20:26 but not 'skindir' 2014-02-28 18:38:22 vkrishn: hey 2014-02-28 18:38:41 ok, you WANT the /usr path 2014-02-28 18:38:55 yes 2014-02-28 18:42:02 ACTION is checking to see what we currently have 2014-02-28 18:42:20 may I ask why you need that dir? 2014-02-28 18:42:40 conditional loading of css file 2014-02-28 18:43:00 based on controller/action 2014-02-28 18:43:10 wouldn't that need the www path? 2014-02-28 18:43:28 you want it loaded by the web client or by the haserl script? 2014-02-28 18:43:28 yes , but need to check first its availability 2014-02-28 18:43:41 on disk 2014-02-28 18:44:32 in general, I have put the link into the HTML whether or not it exists, and allow the client to request it 2014-02-28 18:44:44 if it isn't there, it fails 2014-02-28 18:47:07 eg, I could add file interfaces.css in skin dir, then check if available then create www url and print 2014-02-28 18:48:15 understood 2014-02-28 18:52:52 ACTION had to jump through some hoops in another spot in the code to determine the /usr/share/acf/app/ path from page_info.viewfile 2014-02-28 18:54:08 so, it probably makes sense to add something to page_info to give you /usr/share/acf/app, /usr/share/acf/www/skin, ... 2014-02-28 18:54:51 ACTION might just provide /usr/share/acf/ and let you calculate the skindir 2014-02-28 18:56:46 :) 2014-02-28 18:58:09 also have not checked why, but seems js code in interfaces-create-html.lsp may have some issue with jqmobile framework 2014-02-28 19:01:06 seems to create issue when you try to convert the select menu to jqmobile style menu 2014-02-28 19:02:18 :( 2014-02-28 19:03:03 the jqmoble skin interface would not but fun if the forms not get converted too 2014-02-28 19:07:07 agreed 2014-02-28 19:11:46 vkrishn: I decided that providing /usr/share/acf wouldn't work, because wwwdir and appdir don't have to be at the same path 2014-02-28 19:12:15 so, I will provide page_info.wwwdir that points to /usr/share/acf/www 2014-02-28 19:12:33 please remember to use wwwdir..wwwprefix..skin 2014-02-28 19:13:21 ok 2014-02-28 19:17:25 http://git.alpinelinux.org/cgit/acf-core/commit/?id=e3204e0b67de28068b777aa999305ecec0bc5e7c 2014-02-28 19:20:27 thanks 2014-02-28 19:22:05 vkrishn: http://validator.w3.org/ doesn't complain about the unclosed checkbox input 2014-02-28 19:22:37 strict mode ? or transitional ? 2014-02-28 19:23:08 are you sure it needs to be closed 2014-02-28 19:26:12 In HTML, the tag has no end tag. 2014-02-28 19:26:12 In XHTML, the tag must be properly closed, like this . 2014-02-28 19:27:41 ACTION has been validating with HTML5 (experimental) 2014-02-28 19:28:16 vkrishn: I would like to talk to you more about the label you want, but I have to walk the dog before there is an accident 2014-02-28 19:28:20 back in a few 2014-02-28 19:28:41 'Line 14, Column 70: end tag for "input" omitted, but OMITTAG NO was specified' 2014-02-28 19:29:05 We are wanting to clear XHTML trans or strict 2014-02-28 19:30:26 eg https://dpaste.de/ejjf for reproducing the error 2014-02-28 19:36:49 html5-experimental passes ok for both 2014-02-28 19:49:05 ACTION was shooting for HTML5, not XHTML 2014-02-28 19:49:31 is there a benefit to passing XHTML? 2014-02-28 19:50:30 jqmobile mentions HTML5, but not XHTML 2014-02-28 19:53:00 html5-experimental passes both, and could not find any reference to closing on www.w3.org/TR/2014/CR-html5-20140204/forms.html 2014-02-28 19:53:15 examples given there do not close 2014-02-28 19:53:30 I would go for closing 2014-02-28 19:54:34 quite a few complaints on jqmobile forums about failures on XHTML pages because jqmobile isn't XHTML compliant 2014-02-28 19:54:44 also some bug fixes for this made it into 1.4.1 2014-02-28 19:56:54 since this would require searching the entire tree for views with input tags, I do not have sufficient motivation to change it :) 2014-02-28 19:57:20 now I'm annoyed :/ 2014-02-28 19:57:49 if I come across on w3c html5 docs of any mention of this, I would follow that 2014-02-28 19:58:45 if not follow a well formed element 2014-02-28 19:59:40 understood, and makes sense 2014-02-28 19:59:56 but it goes to my long list of things to do as of now 2014-02-28 20:01:02 vkrishn: now, please help me understand what you want with the label tag 2014-02-28 20:01:20 can label be a priority, with that done, I can almost finish the FirstCut to jqmobile skin 2014-02-28 20:01:40 right now, most inputs have a

with a label 2014-02-28 20:01:59 the checkbox behave correct with jqmobile framework, so do other form elements styling 2014-02-28 20:02:03 are you looking to change that or add something in addition? 2014-02-28 20:03:30 the text in

would be provided by the label so it may be redundant 2014-02-28 20:05:08 but then

can wrap it 2014-02-28 20:07:12 does jquery mobile make use of