2014-09-01 06:51:27 morning 2014-09-01 06:52:11 fabled: should the asterisk group be asterisk? i only see guid 105. 2014-09-01 06:55:02 hi 2014-09-01 06:56:01 clandmeter, should be asterisk. 2014-09-01 06:56:33 fabled: ok. i think something is wrong in 2.7 (dont know about 3/edge) 2014-09-01 06:58:52 fabled: i think if you add a system user, it doesnt create a group be default? 2014-09-01 06:59:30 possible 2014-09-01 07:06:52 morning 2014-09-01 07:07:36 kaniini: i think you were right about redhat wants unify linuxes (kill distros) to simplify support: http://0pointer.net/blog/revisiting-how-we-put-together-linux-systems.html 2014-09-01 07:07:41 they want kill package managers 2014-09-01 07:07:57 and the traditional distro approach 2014-09-01 07:10:40 what will come? 2014-09-01 07:11:15 closed source, propritary things 2014-09-01 07:11:30 i think they want get to the point where all distros are so simlar that there will not really be any point in making a distro 2014-09-01 07:11:50 they want company X able to ship binary-only that runs everywhere 2014-09-01 07:12:06 yup 2014-09-01 07:13:37 that also means that there is no point in supporting alternative libc 2014-09-01 07:14:20 thats probably why they say: "all alternative libcs must mimic glibc" 2014-09-01 07:14:21 you mean like musl/uclibc ? 2014-09-01 07:14:25 yup 2014-09-01 07:14:29 i think they dont like that 2014-09-01 07:14:33 they dont want that 2014-09-01 07:14:49 not quite correct, alternate libc's are good to go 2014-09-01 07:14:52 well, they are ok with it as long as it behaves exactly likg glibc... 2014-09-01 07:15:10 as long as they stick to POSIX 2014-09-01 07:15:19 they dont care about posix 2014-09-01 07:15:20 or any standard 2014-09-01 07:15:24 they care about GNU libc 2014-09-01 07:15:33 they want make their own standard 2014-09-01 07:15:41 there have been flames about that 2014-09-01 07:16:04 people have tried make systemd posix 2014-09-01 07:16:04 oh, I thought musl liked POSIX 2014-09-01 07:16:36 and the response have been: We dont care about posix. if you want systemd support your libc must mimic GNU libc 2014-09-01 07:16:45 *we* like posix 2014-09-01 07:17:01 musl like posix 2014-09-01 07:17:10 then nice 2014-09-01 07:17:22 POSIX is a common play ground that makes it possible that the same app works on multiple systems 2014-09-01 07:17:45 thats why you can have same app on both bsd and linux and osx 2014-09-01 07:17:47 yes, that would make alternate things float and alive 2014-09-01 07:18:04 new thinks are like standard+improvement 2014-09-01 07:18:06 i think lennart is not happy with posix because he does not control it 2014-09-01 07:18:08 things* 2014-09-01 07:18:13 lennart wants make his own standard 2014-09-01 07:18:20 so be it 2014-09-01 07:18:27 and he is trying make people follow it 2014-09-01 07:18:34 no issues lets see what they come up with 2014-09-01 07:19:00 if you want systemd, you will have no option than follow whatever they dictate 2014-09-01 07:19:31 but we are not going for systemd for atleast 2yrs ? 2014-09-01 07:19:56 i dont think we ever will go for systemd 2014-09-01 07:19:58 we are not going for systemd anytime 2014-09-01 07:20:02 the problem is for small distros (like alpine) to work around all that crap with limited resources. 2014-09-01 07:20:04 they don't care about code portability 2014-09-01 07:20:24 nice then 2014-09-01 07:20:29 barthalion: thats what i am trying to say, they dont care about posix 2014-09-01 07:21:27 occasionally its ok if big opensource distros try to push something, be it standard 2014-09-01 07:21:42 if systemd gets standardized (as a part of posix) then maybe... 2014-09-01 07:21:42 it plays ok with economics 2014-09-01 07:23:30 redhat had gone down in last couple of years, nice to see something new 2014-09-01 07:24:02 but that does not mean it cannot be criticized ;) 2014-09-01 07:25:55 found this, performancecopilot.org , seems nice 2014-09-01 07:26:32 there is a good comment on hacker news 2014-09-01 07:26:38 "I actually like their vision - I just wish my linux boxes were not part of it." 2014-09-01 07:27:42 :) 2014-09-01 07:28:11 was looking a new brix's lauched by gigabyte both amd/intel 2014-09-01 07:28:19 seems nice to run AL on it 2014-09-01 07:28:37 they compete with NUC's 2014-09-01 07:30:42 I wish widi and wifi-direct could be implemented soonish 2014-09-01 07:33:29 whats the future vision of alpine? 2014-09-01 07:35:47 to start with "small simple secure" 2014-09-01 07:36:15 so, no systemd? 2014-09-01 07:37:01 I think ncopa already pointed that it would not compile with musl 2014-09-01 07:38:23 vkrishn: which brix? 2014-09-01 07:38:45 http://arstechnica.com/gadgets/2014/04/review-gigabytes-amd-brix-gives-intels-mini-pc-a-run-for-its-money/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+arstechnica%2Findex+%28Ars+Technica+-+All+content%29 2014-09-01 07:39:13 I add to my long term wishlist 2014-09-01 07:39:40 :) 2014-09-01 07:40:11 i have 2 brix at home. I use one as workstation. 2014-09-01 07:40:37 :) 2014-09-01 07:40:48 havent seen that amd one yet. 2014-09-01 07:41:51 I want to build/buy a cheap iscsi, do you know something? 2014-09-01 07:41:58 only the big one, but it makes too much sound for me. 2014-09-01 07:42:26 shafire: hp microserver? 2014-09-01 07:43:15 Looks good, I will check that out 2014-09-01 07:43:31 i like mine 2014-09-01 07:43:36 they are cheap. 2014-09-01 07:43:42 How much did you invest? 2014-09-01 07:43:44 well build 2014-09-01 07:44:08 mine is the ..40.. one 2014-09-01 07:44:15 dont remember exact model name 2014-09-01 07:44:59 i think i spend less then 200 euro 2014-09-01 07:45:16 ive seen them really cheap in uk with some promo code. 2014-09-01 07:46:54 OK, thanks :-) 2014-09-01 07:49:35 clandmeter: http://www.amazon.de/gp/product/B00AHQUX86 2014-09-01 07:50:39 which OS do you use? 2014-09-01 08:00:43 someone using open-iscsi here without problems? 2014-09-01 08:01:09 shafire: which os would you think? ;) 2014-09-01 08:01:25 clandmeter: alpine? freenas? nas4free? :P 2014-09-01 08:01:35 alpine of course 2014-09-01 08:01:51 currently v2.7 with lxc 2014-09-01 08:03:48 still v2.7 due to nfs, although i should try if 3.x works now. 2014-09-01 08:03:59 did you try iscsi too? 2014-09-01 08:04:45 not on this box 2014-09-01 08:04:55 but at office i use alpine iscsi 2014-09-01 08:05:03 but that one is older 2014-09-01 08:05:20 its running scst 2014-09-01 08:06:54 since then, another iscsi implemetnation went mainline so we dont support it anymore on alpine. 2014-09-01 08:07:35 i think we support lio now. 2014-09-01 08:08:02 lio? 2014-09-01 08:08:10 I will check it out 2014-09-01 08:08:33 I can not find this package 2014-09-01 08:08:38 i got lost with all the names 2014-09-01 08:08:54 the userland tools is targetcli i guess 2014-09-01 08:09:16 shafire: http://linux-iscsi.org/wiki/Targetcli 2014-09-01 08:09:19 if you have access to that machine, could you check, if the cpu has the aes flag? :) 2014-09-01 08:09:33 i dont have that cpu 2014-09-01 08:09:38 mine is older 2014-09-01 08:10:55 cat /proc/cpuinfo |grep aes 2014-09-01 08:11:00 nothing. 2014-09-01 08:12:13 ok, thanks 2014-09-02 06:40:52 hi 2014-09-02 06:40:55 good morning 2014-09-02 06:46:51 morning 2014-09-02 07:22:22 someone know cheap ssl wildcard certs?:) 2014-09-02 07:24:24 i think kaniini mention some before, but i dont remember. 2014-09-02 07:27:16 months ago, we had this alphassl wildcard cert for 5$ (5 years) 2014-09-02 07:27:23 apk add lvm2 ERROR: unsatisfiable constraints: madm (missing): required by: world[madm] 2014-09-02 07:27:43 is madm a misspelling of mdadm? 2014-09-02 07:28:24 could be 2014-09-02 07:29:21 lvm2 doesnt depends on mdadm 2014-09-02 07:30:03 i have this weird lxc issue. 2014-09-02 07:30:37 after changing interfaces to static from dhcp and restarting them, after some time the interfaces switches back to its dhcp value again. 2014-09-02 07:31:19 ncopa: i guess dhcpcd is still running in the background taking control of the interface again? 2014-09-02 07:31:55 sounds like that yes 2014-09-02 07:32:04 check if dhcp deamon has exited 2014-09-02 07:32:37 why can't I install lvm2 then? 2014-09-02 07:33:00 i now restarted the containers. cause the resulting ip change is a little bit troublesome. 2014-09-02 07:36:56 shafire: seems you have madm in world? 2014-09-02 07:37:04 can you apk del madm? 2014-09-02 07:37:24 it killed mdadm 2014-09-02 07:37:31 apk del madm (1/1) Purging mdadm (3.3-r1) Executing busybox-1.22.1-r9.trigger OK: 89 MiB in 43 packages 2014-09-02 07:40:15 just add it again if you need it. 2014-09-02 07:41:41 now it worked 2014-09-02 07:41:46 why so complicated? 2014-09-02 07:42:26 so I need to have in /etc/apk/world first lvm2 then mdadm? 2014-09-02 07:42:31 or it will not install the packages? 2014-09-02 07:42:35 after reboot 2014-09-02 07:42:41 somehow you have mdadm in world misspelled 2014-09-02 07:42:51 that shoudlnt happen 2014-09-02 07:43:05 so your system is probably not bitperfect :) 2014-09-02 07:43:36 I have it misspelled? how can this happen 2014-09-02 07:43:52 I did apk add madm and it installed mdadm? 2014-09-02 07:43:52 i dont know. maybe fabled knows. 2014-09-02 07:44:05 nah i dont think thats possible 2014-09-02 07:44:16 uh 2014-09-02 07:44:24 someone typoe 'madm' 2014-09-02 07:44:31 not "m*d*adm" 2014-09-02 07:44:44 see original error 2014-09-02 07:45:14 not sure who would depend on madm 2014-09-02 07:45:16 how can madm refer to mdadm? 2014-09-02 07:45:24 why would it? 2014-09-02 07:45:34 apk del madm (1/1) Purging mdadm (3.3-r1) Executing busybox-1.22.1-r9.trigger OK: 89 MiB in 43 packages 2014-09-02 07:45:54 sounds like someone corrupted world by editing it hand? 2014-09-02 07:46:00 accidentally deleting a letter in it 2014-09-02 07:46:07 "apk del madm" fixed world 2014-09-02 07:46:16 I see, okay 2014-09-02 07:46:17 and because the mdadm was mangled; it got deleted 2014-09-02 07:46:26 ah ok 2014-09-02 07:46:38 that makes sence 2014-09-02 07:46:40 so it was my fault :) 2014-09-02 07:46:47 shafire: did you edit world? 2014-09-02 07:47:11 yes 2014-09-02 07:47:18 duh :) 2014-09-02 07:47:46 and then you question why apk fails... :) 2014-09-02 07:48:10 :p 2014-09-02 07:53:09 What do you think about singularity? 2014-09-02 07:53:36 will you be in the group for or against singularity? 2014-09-02 07:57:52 ? 2014-09-02 08:01:17 barthalion: http://www.youtube.com/watch?v=7GcL3a4WK6M 2014-09-02 08:01:26 watch this if you do not know singularity :) 2014-09-02 08:03:13 this is my biggest fear: http://www.youtube.com/watch?feature=player_detailpage&v=7GcL3a4WK6M#t=3213 2014-09-02 08:11:12 looks like he got himself a smartwatch ;-) 2014-09-02 08:18:44 yeah 2014-09-02 09:14:13 someone using supervisord? 2014-09-02 09:37:51 morning 2014-09-02 10:01:55 #3107 2014-09-02 10:02:06 2 months old :o 2014-09-02 10:03:31 shafire: it does not really introduce any big problem 2014-09-02 10:03:41 I hope :) 2014-09-02 10:04:52 it might cause issues if you have out of memory situation and kernel wants swap out lvm binary/libs 2014-09-02 10:05:08 they want make lvm never swap out 2014-09-02 10:05:14 that is what fails 2014-09-02 10:06:51 ok 2014-09-02 12:03:59 https://www.youtube.com/watch?v=s9FUgzAJun4 2014-09-02 12:23:09 whats different between the normal and xen image? 2014-09-02 12:23:19 just different kernel to boot? 2014-09-02 12:25:54 So, can I switch from standard to xen without any problems? (running from tmpfs and existing apkovl 2014-09-02 12:59:53 shafire: most probably 2014-09-02 13:00:00 you can check alpine-iso for diferences 2014-09-02 13:00:14 I checked /etc/apk/world and /etc/modules 2014-09-02 13:00:16 no difference 2014-09-02 13:00:48 alpine-iso builds the iso images. so if there something special, it should be listed. 2014-09-02 13:01:16 ok 2014-09-02 14:07:24 sigh, freeradius dev suffer from the NIH syndrome... 2014-09-02 14:07:38 i spend a couple of days to provide a proper patch 2014-09-02 14:07:51 and then then go and make their own variant 2014-09-02 14:08:01 with fixed size problem 2014-09-02 14:17:54 where can I find a init.d template script 2014-09-02 14:19:03 http://wiki.alpinelinux.org/wiki/Writing_Init_Scripts 2014-09-02 14:22:44 that was easy 2014-09-02 14:28:13 will we have xen 4.4 on armhf? 2014-09-02 14:29:35 q 2014-09-02 15:41:21 shafire: ipxcore used to have them for $40 2014-09-02 16:05:11 http://alpinelinux.org/forum/general-discussion/run-startup-script-after-packages-installed 2014-09-02 16:05:13 what does he mean? 2014-09-02 16:32:33 debian has that feature 2014-09-02 16:32:40 where the post-install runs the init script 2014-09-02 16:32:55 oh 2014-09-02 16:32:58 no 2014-09-02 18:48:11 htop is ok - I needed to "apk upgrade --available" after upgrading to musl 2014-09-02 19:01:41 shafire: I guess ther are bigger problems around me to worry about 2014-09-02 19:01:46 maybe not bigger, let's say closer 2014-09-03 06:15:27 i need to turn off arm builder for a while 2014-09-03 07:37:23 Morning 2014-09-03 13:19:39 files from v2.5.4-339-gb8346d0 uploaded 2014-09-04 07:57:21 morning 2014-09-04 08:00:00 still having no luck with X on this Dell D620 (intel 945GM) 2014-09-04 08:07:33 where is KMS configured in alpine linux? 2014-09-04 08:26:11 i think /etc/modprobe.d/kms.conf 2014-09-04 08:43:12 Does kms.conf actually load modules? i915 is in there but nothing happens on boot 2014-09-04 08:43:34 it should be autoloaded 2014-09-04 08:43:41 if i load it manually, it sets my screen size 2014-09-04 08:43:50 hm 2014-09-04 08:43:55 on the console - no X yet 2014-09-04 08:43:58 try add it to /etc/modules then 2014-09-04 08:46:56 I have just done that, and now I have a black screen after some boot messages, once init loads the module 2014-09-04 08:49:59 try add fbdev there too 2014-09-04 08:52:38 instead? 2014-09-05 08:36:00 huh 2014-09-07 07:55:53 I upgraded from alpine 2.6 branch to 2.7 2014-09-07 07:56:03 now I can't log into my samba shares and I have this in syslog 2014-09-07 07:56:04 Sep 7 07:54:38 obsidian daemon.err smbd[21338]: [2014/09/07 07:54:38.822331, 0] ../lib/util/become_daemon.c:136(daemon_ready) 2014-09-08 04:33:43 I need a build of Alpine armhf targeted to a arm5tel CPU (PogoPlug v4). Can I get some links to directions on how to setup an Alpine build environment for cross-compilation? 2014-09-08 13:13:25 last is broken too :( 2014-09-08 15:22:26 mh 2014-09-08 15:22:39 whats the init script, which overlays the usb content? 2014-09-08 15:45:49 hello all 2014-09-08 15:46:35 if some package need libintl.h how to resolve this with musl lib? 2014-09-08 15:56:30 nvm - found it in gettext pkg, sry :) 2014-09-08 15:57:00 how can I let an init script run as another user? 2014-09-08 15:57:12 or set env variables there? 2014-09-08 16:00:50 try daemonize 2014-09-08 16:38:03 luckly, someone packed supervisor 2014-09-08 16:38:05 I am using this now 2014-09-08 16:38:45 shafire: ur welcome 2014-09-08 16:38:47 :] 2014-09-08 16:39:10 shafire: start-stop-daemon also let's you specify uid and gid w/ --user 2014-09-08 16:39:41 pnutzh4x0r: I installed supervisor with easy_install 2014-09-08 16:39:47 on tmpfs 2014-09-08 16:39:57 oh 2014-09-08 16:39:58 its very hard to get it run in this env 2014-09-08 16:40:04 on* 2014-09-08 16:40:22 and the init script should start supervisor 2014-09-08 16:40:33 but now with the package, problem is gone :) 2014-09-08 16:41:29 so thanks pnutzh4x0r 2014-09-08 16:41:52 no problem 2014-09-09 01:37:10 how can I build all the alpine-sdk packages from scratch 2014-09-09 01:43:54 can I use abuild to cross compile? 2014-09-09 01:53:02 yes. fabled has a solution which is more elegant than the hard way. ask him when he shows up. 2014-09-09 01:53:41 kaniini: looking forward to it.... 2014-09-09 03:47:27 kaniini: how often is fabled online, daily? 2014-09-09 03:50:10 he should be on in a few hours. 2014-09-09 03:53:30 don't think I'll be online in a few hours. is there a way to have irc deliver a message to him when he logs onto irc? 2014-09-09 03:54:06 somebody know about support fts.h in musl ? 2014-09-09 03:57:23 there used to be, fts-dev 2014-09-09 03:58:28 did not find such package, did i miss something? 2014-09-09 03:58:55 it probably needs to be rebuilt again. we added fts support to uclibc. 2014-09-09 03:59:06 i will talk to musl guys about adding it there 2014-09-09 05:05:58 pogoplug: there is some freenode service, I don't remember now 2014-09-09 05:06:03 memosomething 2014-09-09 05:31:33 pogoplug: hi 2014-09-09 10:03:29 ncopa, found another dir/file watch util http://puszcza.gnu.org.ua/projects/direvent 2014-09-09 10:03:42 add it to http://bugs.alpinelinux.org/issues/2443 or new request ? 2014-09-09 10:04:17 how about inotifyd from busybox? 2014-09-09 10:04:50 I tried, but found was it did not notify if changes where quick 2014-09-09 10:04:54 sometimes 2014-09-09 10:05:57 but then that may be issue with other similar utils, yet to test 2014-09-09 10:07:28 nice fs article, http://www.phoronix.com/vr.php?view=20877 2014-09-09 10:08:22 #2285, hope to get this 2014-09-09 10:09:49 this looks interesting http://git.gnu.org.ua/cgit 2014-09-09 10:36:01 http://lwn.net/Articles/605128/ has some nice article, see comments, "btrfs has change log stgreaming" 2014-09-09 10:36:14 streaming* 2014-09-09 10:37:10 http://marc.merlins.org/perso/btrfs/post_2014-05-19_Btrfs-diff-Between-Snapshots.html 2014-09-09 10:49:01 ncopa, anything similar to https://www.gitorious.org/obexpushd/ 2014-09-09 10:49:16 for receiving files from mobile ? 2014-09-09 11:08:30 apk search obex 2014-09-09 11:29:32 trying obexftp 2014-09-09 14:30:46 kaniini: did I miss fabled? 2014-09-09 14:35:39 fabled: I'm trying to cross-compile from x86_64 to armv5tel using http://ellcc.org/. Have any advice on how I can build my cross compiling environment? Should I start with an Alpine edge environment with alpine-sdk and then override some enviroment variables? 2014-09-09 14:37:08 pogoplug, i have no experience with ellcc. but yes, take alpine edge, and modify scripts as needed. the stuff i used for cross-building is at http://dev.alpinelinux.org/~tteras/bootstrap/ (it may or may not work) 2014-09-09 14:43:15 fabled: that looks a lot easier than reading apk-tools/ABUILD.conf and manually building its dependencies then it. 2014-09-09 14:45:31 thanks! 2014-09-09 15:15:10 somebody know how can i compile lisp file in alpine ? 2014-09-09 15:15:19 without emacs? 2014-09-09 16:40:37 j 2014-09-10 00:59:54 how do I get abuild to run as root? 2014-09-10 01:27:40 How about insight on this abuild message: >>> ERROR: gnupg-pkcs11: download is missing in checksums 2014-09-10 01:33:21 pogoplug: i think that means that "download" is missing a checksum entry 2014-09-10 01:33:28 id check the url or source u are using 2014-09-10 01:35:01 alternatively did u generate the checksum? 2014-09-10 01:35:22 (i.e. abuild checksum) 2014-09-10 01:39:23 pnutzh4x0r: turns out it was a bad link to sourceforge.net ending with /download instead of going to downloads.sf.net//.tgz 2014-09-10 01:39:30 pnutzh4x0r: thanks for the input 2014-09-10 01:40:48 np 2014-09-10 03:21:02 is tteras online here? 2014-09-10 03:41:18 nevermind, that's fabled 2014-09-10 06:24:02 hm 2014-09-10 06:24:34 seems we need CONFIG_RELOCATABLE for the EFI stub suppport now 2014-09-10 06:25:09 scary change to enable that in a stable branch 2014-09-10 11:06:38 faffolter: it seems like py-dateutil installs with wrong permissions 2014-09-10 11:07:43 http://sprunge.us/QPUI 2014-09-10 11:08:05 i also see that fedora reverted back to python-dateutil-1.5: http://pkgs.fedoraproject.org/cgit/python-dateutil.git/log/ 2014-09-10 11:08:11 do you have any idea why they reverted? 2014-09-10 11:08:37 it breaks py-matplotlib 2014-09-10 11:08:40 ACTION is asking his monthly question: any news regarding nfs? 2014-09-10 11:08:59 StarWarsFan: not more than nfs-server should work 2014-09-10 11:09:13 i haent had time to look at the nfs client issue 2014-09-10 11:09:13 woahhhh, really??? 2014-09-10 11:09:21 ah, well 2014-09-10 11:09:23 client 2014-09-10 11:09:46 ok 2014-09-10 11:11:36 ok, will wait what happens. i'm still very busy with planing our own house, so no time for anything else beside the job... 2014-09-10 14:18:03 fabled: can you assist with my cross build? I'm trying to understand the build cross toolchain vs cross bootstrap and what variables to populate it with. I'm building on x86_64 to target armv5te 2014-09-10 14:51:26 fabled: are you online? 2014-09-10 20:15:45 ncopa: i will look into it 2014-09-11 02:36:56 I'm running into a problem trying to build some new packages. I started trying to do a cross compile following fabled's scripts, and now my system seems to only want to build cross compile. I get armv6-alpine-linux-muslgnueabihf in the output of abuild which suggests somehow my environment is still hooked to the cross build. any ideas on how to revert it to x86_64? 2014-09-11 02:38:30 my shell environment is clean of any C{,PP,XX}FLAGS 2014-09-11 02:43:35 check /etc/abuild.conf ? 2014-09-11 02:48:21 CHOST=x86_64-alpine-linux-musl 2014-09-11 02:54:12 there is also ~/.abuild/abuild.conf 2014-09-11 02:54:24 but other than that, i have no idea 2014-09-11 02:55:36 ah, let me check that 2014-09-11 02:59:11 still nothing 2014-09-11 03:17:24 pnutzh4x0r: any other ideas on where an armv6 variable might get stored, perhaps cached in /var/abuild? 2014-09-11 03:20:24 pnutzh4x0r: should abuild -a work? I'm getting illegal option 2014-09-11 03:23:27 pogoplug_: sorry, i dont know too much about the internals of abuild... ive only used it to make a few packages 2014-09-11 03:30:59 pnutzh4x0r: I appreciate your comments, anything that might jog the mind to think of something new/useful 2014-09-11 03:43:01 the problem is that most of the alpine devs are european, so this is not a good time to ask for help 2014-09-11 03:43:05 at least, i think they are 2014-09-11 06:03:03 is it ok that there is no init.d in source files in APKBUILD of backuppc ? 2014-09-11 06:03:44 and also in backuppc apk package, there is also no init.d script 2014-09-11 06:04:47 k0r10n: http://git.alpinelinux.org/cgit/aports/tree/main/backuppc/backuppc.initd 2014-09-11 06:05:29 looks like backuppc needs a cleanup 2014-09-11 06:06:58 i mean in APKBUILD, source string - there is only config.pl , but no pre-install or post-install 2014-09-11 06:07:06 also 2014-09-11 06:14:49 yeah, i think it needs cleanup 2014-09-11 08:43:01 fabled: hi 2014-09-11 08:43:41 seems like claws-mail segfaults on freebsd too, with utf8 locale (but not de_DE.ISO8859-1) 2014-09-11 08:43:46 when creating vcard 2014-09-11 08:43:57 bactrace from freebsd: http://pastebin.com/wGj2RQkp 2014-09-11 08:44:04 i think it is different from our issue 2014-09-11 08:57:15 ncopa, yes, looks different 2014-09-11 08:57:41 the crashes i see always are in free() in places indicating earlier heap corruption 2014-09-11 08:58:25 likely due to a race... 2014-09-11 08:58:40 segfault happens much more seldom when running in gdb 2014-09-11 08:58:46 but it did happen eventually 2014-09-11 08:59:22 i saw a glib warning about refcount 2014-09-11 08:59:40 might be that 2 threads unref same object 2014-09-11 08:59:48 or similar 2014-09-11 10:46:13 are there some news about fts in musl? 2014-09-11 13:01:48 k0r10n: what needs fts.h? 2014-09-11 14:09:49 ncopa: i'm trying to build elfutils 2014-09-11 14:13:38 we used to have a package with fts 2014-09-11 14:13:48 git log -- main/fts 2014-09-11 14:13:58 git revert 01907687add4bd2c41053f1314ad423d6c9c21fe 2014-09-11 14:14:02 should bring it back 2014-09-11 14:15:01 yes i saw old version fts in uclibc version 2014-09-11 14:16:13 i think it builds at least 2014-09-11 14:16:17 i can add it back to testing 2014-09-11 14:17:37 would be great 2014-09-11 14:17:52 tyvm 2014-09-11 14:25:46 morning 2014-09-11 14:25:57 anyone tried zfs support yet? 2014-09-11 15:00:45 fabled: I messed up my abuild envrionment somehow. I run abuild as a non-root user, and I'm getting the following in the output.>>> pkcs11-helper: Unpacking /var/cache/distfiles/pkcs11-helper-1.11.tar.bz2... armv6-alpine-linux-muslgnueabihf 2014-09-11 15:01:30 any idea where the armv6 string might be coming from? I'm building a new package on an x86_64 that I was also planning to run a cross compilation from x86_64 -> armv5te. 2014-09-11 15:01:54 fabled: I started with your create-cross and cross-bootstrap scripts,, but I thought they didn't work for me, and now the system seems to be wonky. 2014-09-11 16:15:49 GISTER p0g0 pagerc@gmail.com 2014-09-11 16:16:08 haha this is painful 2014-09-11 16:17:47 sorry for spewing my nick registration to all. web irc on cell phone makes typing a pain 2014-09-11 16:23:57 somebody know why GLOB_TILDE is not defined in glob.h ? 2014-09-11 18:15:25 Has anyone proposed moving away from gcc as the default compiler? 2014-09-11 18:23:13 please never switch to systemd 2014-09-11 18:29:56 But please fork their dhcp client as it's supposed to break the spec to acquire an IP faster, like MacOS 2014-09-11 18:30:24 why not put a pull request into udhcpd? 2014-09-11 18:31:05 Perhaps, I don't always suggest smart things 2014-09-11 18:31:06 but i basically hate everything systemd stands for 2014-09-11 18:31:26 because it's not UNIX philosophy, which is "do one thing, but do it right/good" 2014-09-11 18:31:28 It does a lot of good things 2014-09-11 18:31:57 that's the issue, it's overloaded with crap that nobody wants in a production system 2014-09-11 18:32:02 Such as async blocking by owning all file descriptors that might block 2014-09-11 18:32:30 we advised all clients to migrate to non-systemd installations 2014-09-11 18:32:32 It has some stuff that's very useful. And a lot of cruft 2014-09-11 18:32:50 worked, we got a lot more gentoo and alpine dom0's (with ubuntu domUs and systemd) 2014-09-11 18:32:56 That's going to be harder to support in the future 2014-09-11 18:33:06 Ugh don't Xen 2014-09-11 18:33:16 GenodeOS 2014-09-11 18:33:18 we do xen since v2 2014-09-11 18:33:36 Xen is bad design 2014-09-11 18:33:45 but KVM is better? :> 2014-09-11 18:34:21 genode looks like it's made for desktops 2014-09-11 18:34:25 we only administrate servers 2014-09-11 18:34:32 LXC/Docker/systemd -nspawn 2014-09-11 18:34:46 yeah never in a gazillion years i'm falling for that one ;) 2014-09-11 18:35:01 btw our freebsd jails setups went up too 2014-09-11 18:35:07 GenodeOS is L4 microkernel with Linux l4env 2014-09-11 18:35:21 imho xen is great if you know what you're doing 2014-09-11 18:35:28 but so is all technology 2014-09-11 18:35:33 i've heard people being happy with php 2014-09-11 18:35:35 Not falling for what? LXC? 2014-09-11 18:35:38 so mysteries do exist 2014-09-11 18:35:51 both, docker and LXC 2014-09-11 18:35:56 i don't like both 2014-09-11 18:36:16 Hmm, LXC is nice 2014-09-11 18:36:18 linux containers are a fault that was never cleaned up (imho) 2014-09-11 18:36:23 hrhr ok 2014-09-11 18:36:27 so we're off on philosophy 2014-09-11 18:36:29 thats fine 2014-09-11 18:36:43 But it's not enforced security if there's kernel exploits 2014-09-11 18:37:00 well grsec-xen is just great for that 2014-09-11 18:37:09 My issue with Xen is drivers from dom0 2014-09-11 18:37:10 with PaX of course 2014-09-11 18:37:27 well we're keeping to a quite specific set of servers that work very well 2014-09-11 18:37:40 except for them sometimes requiring new firmware drivers in grsec.gz 2014-09-11 18:38:34 Xen works, but the dom0 integration was always it's architectural flaw 2014-09-11 18:38:53 It might be better. I haven't looked at it since 2004 2014-09-11 18:39:16 pogoplug: moving to clang is possible one day 2014-09-11 18:39:38 fbettag: while to systemd not, as they don't care about non-glibc standard libraries 2014-09-11 18:41:27 barthalion: *whew* 2014-09-11 18:54:35 i'm so happy with alpine as it is 2014-09-11 18:54:41 alpine-iso is awesome 2014-09-11 18:55:08 all our images are dd flashable to client servers thus we could automate it and now do not need to invoice our clients for setups 2014-09-11 18:55:12 it's SO awesome 2014-09-11 21:31:04 ncopa: when can we expect the aport for zfs to appear in testing? ツ 2014-09-11 21:37:41 the biggest problem is that we need to rework the patch from tirpc to something that allows linking to the zfs license :/ 2014-09-11 21:38:31 but technically it works if i build it myself? 2014-09-11 21:46:44 yes. 2014-09-11 21:47:10 but it needs some cleanups. 2014-09-11 21:56:36 hmpf ok 2014-09-11 21:56:39 too bad 2014-09-11 21:56:50 i am in the dc next week 2014-09-11 21:57:00 wouldve tried it on a new storage 2014-09-11 22:04:35 ERROR: unsatisfiable constraints: 2014-09-11 22:04:35 .makedepends-spl-grsec-0: 2014-09-11 22:04:35 masked in: cache 2014-09-11 22:04:36 satisfies: world[.makedepends-spl-grsec] 2014-09-11 22:04:38 meh 2014-09-11 22:13:35 there is a spl-grsec patched up. 2014-09-11 23:26:26 where is "update_config_sub" defined for APKBUILD files? 2014-09-11 23:38:59 is it safe to upgrade to edge right now? anything broken? 2014-09-11 23:42:23 wow, I totally just found this mailing list called alpine-devel where this guy Natanael Copa and another guy Timo Teras are discussing the intracies of getting abuild to cross-build to arm and specifying the libc to use explicilty 2014-09-12 02:56:35 ncopa: any idea why update_config_sub was returning info about armv6 cross build and failing an apk build on an x86_64 host? 2014-09-12 03:18:28 I try running createcross-toolchain and get the following output: binutils-x86_64-alpine-linux-musl (missing): 2014-09-12 03:18:38 any takers on how to resolve that? 2014-09-12 08:14:07 can anoyne tell me where i can find /usr/bin/truncate? 2014-09-12 08:15:02 coreutils 2014-09-12 08:15:03 gnaa 2014-09-12 08:33:19 ncopa: zlib-dev is missing from make_depends on zfs-grsec 2014-09-12 08:34:00 also coreutils 2014-09-12 08:45:38 fbettag: thank you, and i can't understand why so many headers missing...:) 2014-09-12 08:45:47 fbettag: i dont think it builds at all... 2014-09-12 08:45:53 fbettag: while compiling 2014-09-12 08:46:00 ncopa: i also adjusted configure.ac 2014-09-12 08:46:18 also there is a typedef missing 2014-09-12 08:46:31 i think there are atleast 2 pull requests for musl support in zfs github 2014-09-12 08:46:37 zxprrg ___builtin_va_list ___gnuc_va_list; 2014-09-12 08:46:44 fbettag:va_list 2014-09-12 08:46:46 oh yes 2014-09-12 08:46:51 the first should've been typedef 2014-09-12 08:47:06 needs to be added to lib/libspl/petexecname.c 2014-09-12 08:47:14 and gethrtime.c 2014-09-12 08:49:24 ncopa: maybe i can fix it and we have it as patches 2014-09-12 08:51:33 https://github.com/zfsonlinux/zfs/pull/2605 and https://github.com/zfsonlinux/zfs/pull/2604 2014-09-12 08:52:12 i better try their git then 2014-09-12 08:52:54 but the first does not look good 2014-09-12 08:53:01 2605 2014-09-12 08:53:38 someone should clean them up properly and make a new pull request 2014-09-12 08:55:50 "The fact that porting the code to MUSL highlighted this is a credit to the benefits of writing portable code." 2014-09-12 08:55:52 heh 2014-09-12 08:56:08 hrhr 2014-09-12 08:56:12 they found bugs while trying to port to musl :) 2014-09-12 08:59:10 fbettag: you might spare some time and can find a musl-patch in my repo, this is the basis for the pullrequests. also you can try to clean up the pull request.... 2014-09-12 08:59:25 ncopa: that's often the case when you try out other libs/compilers 2014-09-12 08:59:29 github/stef/aports 2014-09-12 08:59:38 ncopa: freebsd found lots of bugs while switching to clang 2014-09-12 09:00:56 you are doing redundant work now 2014-09-12 09:02:43 f15RyWpgGTO6: allright 2014-09-12 09:03:18 f15RyWpgGTO6: which one is yours? 2014-09-12 09:03:23 all of them. ;) 2014-09-12 09:03:24 ah 2014-09-12 09:03:28 stef ;) 2014-09-12 09:05:00 f15RyWpgGTO6: well i have the configure.ac fixed so it does the subdirs 2014-09-12 09:05:36 and proper modern init 2014-09-12 09:06:28 huh? 2014-09-12 09:06:55 configure.ac:41: warning: AM_INIT_AUTOMAKE: two- and three-arguments forms are deprecated. 2014-09-12 09:07:04 cmd/mount_zfs/Makefile.am:14: warning: source file '$(top_srcdir)/cmd/mount_zfs/mount_zfs.c' is in a subdirectory, 2014-09-12 09:07:07 cmd/mount_zfs/Makefile.am:14: but option 'subdir-objects' is disabled 2014-09-12 09:07:19 there are a ton of these 2014-09-12 09:07:40 ah, but those are warnings. and i believe unrelated to musl. 2014-09-12 09:07:45 i didn't care about those 2014-09-12 09:08:07 i'll put a pull request for you in 2014-09-12 09:08:13 or you just want those 3 lines pasted? 2014-09-12 09:08:28 (i prefer pasting) 2014-09-12 09:09:03 depends which repo. the zfs or the aports? 2014-09-12 09:09:19 also a patch would be nicer than just 3 lines. 2014-09-12 09:10:24 i think a direct pull request against ZoL would make more sense. as this is unrelated to the musl port. 2014-09-12 09:11:39 kk 2014-09-12 09:13:22 In file included from ../../lib/libspl/xdr.c:33:0: 2014-09-12 09:13:23 ../../lib/libspl/include/rpc/xdr.h:35:26: fatal error: rpc/xdr.h: No such file or directory 2014-09-12 09:13:25 #include_next 2014-09-12 09:13:28 ^ 2014-09-12 09:13:30 compilation terminated. 2014-09-12 09:13:33 Makefile:728: recipe for target 'xdr.lo' failed 2014-09-12 09:15:09 fbettag: xdr.h is in libtiprc-dev 2014-09-12 09:15:25 missing dep then ;) 2014-09-12 09:15:45 fbettag: but this file is in /usr/include/tirpc folder 2014-09-12 09:20:57 ooh? 2014-09-12 09:22:13 any of you tried it in production 2014-09-12 09:22:24 i only used it on freebsd/osol 2014-09-12 09:23:53 i tried only on gentoo 2014-09-12 09:24:10 was it stable or any good? 2014-09-12 09:24:15 btw netconfig.h is missing 2014-09-12 09:24:21 where's that? 2014-09-12 09:24:23 i'm using my own patch reliably 2014-09-12 09:25:06 netconfig.h is in /usr/include/tirpc/netconfig.h 2014-09-12 09:26:37 fbettag: are you talking about my patch and my apkbuild or the official one? 2014-09-12 09:26:45 f15RyWpgGTO6: yours 2014-09-12 09:26:46 ;) 2014-09-12 09:26:50 strange 2014-09-12 09:27:02 i started from a very fresh system 2014-09-12 09:27:10 so i notice all the things that aren't documented ;) 2014-09-12 09:30:11 fbettag: then you've got error in header xdr.h in lib/libspl/include/rpc/xdr.h 2014-09-12 09:30:19 fbettag: i'm stuck on it right now 2014-09-12 09:30:32 yup 2014-09-12 09:32:54 using my patch? 2014-09-12 09:32:58 what are you using? 2014-09-12 09:33:13 git HEAD or 0.6.3? 2014-09-12 09:33:37 head 2014-09-12 09:34:11 using your repo 2014-09-12 09:34:21 in my apkbuild there is libtirpc-dev as a dependency: depends_dev="glib-dev e2fsprogs-dev util-linux-dev libtirpc-dev 2014-09-12 09:34:35 ah i was using ncopa's 2014-09-12 09:34:40 ahaa 2014-09-12 09:34:56 so which of your issues pertains to my patch? 2014-09-12 09:35:25 where do i find your apkbuild? ;) 2014-09-12 09:35:36 also my patch uses 0.6.3 and not head. 2014-09-12 09:35:46 10:59:19 f15RyWpgGTO6 | github/stef/aports 2014-09-12 09:35:48 i'm using your master 2014-09-12 09:36:07 (oh, i'm refering to the version of zfs not the patch version) 2014-09-12 09:49:10 f15RyWpgGTO6: still compiling 2014-09-12 09:50:58 f15RyWpgGTO6: worked! 2014-09-12 10:08:06 f15RyWpgGTO6: thanks a lot! 2014-09-12 10:16:37 fbettag: strange, i'm still getting error in xdr.h 2014-09-12 10:17:13 fbettag: error:expected indetifier or '(' before 'if' 2014-09-12 10:17:39 k0r10n: dont play with sources use github.com/stef/aports 2014-09-12 10:17:42 fbettag: in lib/libspl/include/rpx/xdr.h:64:15 2014-09-12 10:17:43 theres your apkbuild 2014-09-12 10:17:46 which patches and works 2014-09-12 10:22:22 trying out my iso 2014-09-12 10:23:00 fbettag: tyvm, you were right, wrong APKBUILD.... 2014-09-12 10:41:43 f15RyWpgGTO6: any idea where zfs.ko is laying around? 2014-09-12 10:42:53 huh 2014-09-12 10:42:56 hm 2014-09-12 11:10:54 f15RyWpgGTO6: modloop mounted doesnt like placng stuff into modules 2014-09-12 11:12:21 apart from that 2014-09-12 11:12:47 insmod: can't insert '/lib/modules/3.14.17-0-grsec/extra/zfs/zfs.ko': unknown symbol in module, or unknown parameter 2014-09-12 11:22:42 try a couple times 2014-09-12 11:22:53 with the other dpendencies 2014-09-12 11:24:08 depending modules i mean 2014-09-12 11:33:49 kk 2014-09-12 11:54:49 i just noticed that by using tar -u makes it even more awesome to just update the apkovl 2014-09-12 12:34:37 I noticed that tar -u sometimes does not recurse dir well, for changes 2014-09-12 12:36:17 ncopa, iPhone 6 has nfc ;) 2014-09-12 12:48:44 cool, too bad i dont have iphone 2014-09-12 12:48:53 didnt android phones have nfc for a long time? 2014-09-12 12:48:58 nfc sounds like old tech... 2014-09-12 12:49:06 yes 2014-09-12 12:49:26 it lost traction for a while 2014-09-12 12:51:31 will AL benefit from adding nfc ? 2014-09-12 12:52:14 I was thinking someone implementing POS(point of sale) system or ticketing booth... etc using AL ;) 2014-09-12 12:54:26 yet another wishlist, http://www.dailytech.com/ZOTAC+Announces+Diminutive+ZBOX+PI320+pico+/article36458.htm 2014-09-12 12:56:17 but costly, considering some new tablets have have similar features at cheaper rate 2014-09-12 13:43:24 ncopa: it's wired. it seems that the developer of dateutil is maintaining two version with only a small difference in the name. 2014-09-12 13:43:47 ncopa: i guess it would be the easiest way to go back to 1.5 for now 2014-09-12 14:06:23 fabled: can you assist with debugging a cross build with those scripts you pointed me at? 2014-09-12 14:06:38 pogoplug, i'm sort of busy this and next week 2014-09-12 14:06:50 but if you do only cross compiler, you'll need the first script only 2014-09-12 14:07:02 I need cross compile + cross build 2014-09-12 14:08:00 It looks like for the first cross compiler I need an x86_64/x86_64 then for the cross build I need x86_64/armv5te so that I'm specifying the CHOST/CTARGET for the compiler to be the same architecture then for the cross build that I'm running on x86_64 but targetting armv5te, is that about right? 2014-09-12 14:16:39 pogoplug, ? 2014-09-12 14:16:49 there's two scripts 2014-09-12 14:17:05 1. creates a compiler that runs on x86_64, and generates arm code 2014-09-12 14:17:24 2. using the compiler from step 1., it compiles compiler that runs on arm and generates arm code 2014-09-12 14:17:38 oh, humm 2014-09-12 14:17:40 if you want to *only* cross compile (not compile on arm it self) 2014-09-12 14:17:44 you need to run the step 1 2014-09-12 14:17:52 createcross-toolchain.sh 2014-09-12 14:18:02 you install that locally, and try to cross compile then 2014-09-12 14:18:15 ERROR: unsatisfiable constraints: binutils-x86_64-alpine-linux-musl (missing): 2014-09-12 14:18:17 i'm not sure how much of alpine will cimple 2014-09-12 14:18:22 cross-compile* 2014-09-12 14:18:32 we fixed only the core set to be cross compilable 2014-09-12 14:18:39 There's the armhf build you did, right? 2014-09-12 14:18:41 as cross compiling can be tricky with certain packages 2014-09-12 14:18:43 yes 2014-09-12 14:18:52 that's also how we bootstrapped musl 2014-09-12 14:18:55 I'm really just looking for core+select packages 2014-09-12 14:19:12 well 2014-09-12 14:19:24 you can use crossbuild-alpine* script to core cross compiling 2014-09-12 14:20:14 is there an alpine package for this, or just adhoc scripts? I ask because it'd be nice to pull your scripts out of git or something similar. 2014-09-12 14:21:07 btw, thanks for the assistance, really appreciate the pointers 2014-09-12 14:22:45 no package 2014-09-12 14:22:54 and just to triple check, I should be using the scripts from http://dev.alpinelinux.org/~tteras/bootstrap/ 2014-09-12 14:22:56 we try to avoid bootstrapping 2014-09-12 14:23:03 yes, that's what i used 2014-09-12 14:23:12 there's no 'official' repository, git tree, or package 2014-09-12 14:23:18 gotcha 2014-09-12 14:23:50 and it works only with musl 2014-09-12 14:23:56 I used to start with a gentoo stage3 tarball, and rebuild the system from scratch, was hoping something similar here. The big win from alpine is that musl is used by default and there's basic cross compilation support 2014-09-12 14:23:59 bootstrapping glibc/uclibc is pain 2014-09-12 14:24:12 I'm only interested in musl 2014-09-12 14:24:30 I'm hoping to switch from gcc to clang/ellcc, but for now I'll get gcc building it. 2014-09-12 14:24:31 yeah. we don't have gentoo like stage1/2/3 scripting done too often 2014-09-12 14:24:53 If I get a full core packages built for armv5te, is the project interested in hosting them parallel to the armhf? 2014-09-12 14:25:10 i think yes 2014-09-12 14:25:27 i've been thinking doing armv7+/thumb build too 2014-09-12 14:25:47 though i wonder how portable the kernels are 2014-09-12 14:25:54 we probably would need kernel per board on armv5 2014-09-12 14:26:42 when I run create-cross-toolchain, will that break my ability to build x86_64 packages from that abuild environment? 2014-09-12 14:26:49 no 2014-09-12 14:27:03 but it might prevent gcc upgrade 2014-09-12 14:27:11 i think gcc had versioned dependencies 2014-09-12 14:27:12 fabled: I'd have to look at how device tree support works, but might be possible to run a common kernel for the little endian boards 2014-09-12 14:27:31 pogoplug, yes, that's what i'm doing with wandboard+other boards now 2014-09-12 14:27:36 though, rpi needs it's own kernel 2014-09-12 14:27:54 I've got some beagle bone black boards that I'm willing to make ssh accessible for testing 2014-09-12 14:29:52 fabled: thanks for all the input while you're busy and all, I'll see about debugging the binutils-x86_64-alpine-linux-musl missing dependency when running the ./crosscreate-toolchain 2014-09-12 14:30:10 there might have been packaging changes since i used them 2014-09-12 14:30:15 so that might cause the issues 2014-09-12 14:30:32 ok, I'll dig in I guess. 2014-09-12 16:58:25 please git pull git://git.alpinelinux.org/fab/aports 2014-09-12 16:58:35 there are a couple of updates 2014-09-12 16:58:36 thanks 2014-09-12 22:55:04 what's up2date supposed to do? 2014-09-12 22:55:13 localhost:~/aports/main/musl$ abuild up2date >>> ERROR: musl: up2date failed 2014-09-12 23:32:30 pogoplug: fails if there is a new upstream version 2014-09-12 23:36:42 kaniini: does that suggest that aports-git is broken somehow? 2014-09-12 23:46:52 why does abuild APKBUILD reference CPAN/perl? 2014-09-13 19:34:56 would anyone complain if I fixed abuild to run faster by using case statements instead of [ $var == "str" ] comparisons? 2014-09-13 20:01:00 Is /usr/share/abuild/function.sh in readconfig() supposed to be called when sourcing the functions directory, or only when explicitly called via the readconfig function. I ask because I'm trying to run abuild, and its changing me to a directory with a non-existent ./APKBUILD which cause abuild to fail 2014-09-13 20:05:07 is there a source control repo for the abuild project so that I can review the commit history to identify when it "broke"? 2014-09-13 20:06:06 nevermind, found the info via apk info abuild 2014-09-13 20:11:36 $#!@$#@% so....abuild depends on coreutils, readlink -f is broken in busybox such that if its not a link, it returns an empty string... 2014-09-14 01:01:53 kaniini: ever see an error like: "libtool: compile: not configured to build any kind of library" 2014-09-14 04:33:11 any idea when that 'last' bug might be fixed? 2014-09-15 07:07:19 more nfc news 2014-09-15 07:07:21 http://www.nxp.com/news/whats-cooking-in-rd/2034-mbps-nfc.html 2014-09-15 07:07:22 http://www.nxp.com/news/press-releases/2014/08/nxp-releases-complete-passive-solution-for-nfc-interaction-with-electronic-devices.html 2014-09-15 07:07:22 http://www.nxp.com/news/press-releases/2014/02/nxp-expands-nfc-ecosystem-with-android-kitkat-integration.html 2014-09-15 07:07:29 <[xming]> morning all 2014-09-15 07:09:13 <[xming]> trying to migrate some of my VMs to alpine and noticed that it misses some pkgs which I need so I am trying to make some 2014-09-15 07:09:39 <[xming]> stumbled on the 1st pkg already :/ 2014-09-15 07:10:08 <[xming]> conf.h:140:3: error: 'uid_t' does not name a type 2014-09-15 07:10:43 ncopa, spliting #2437 for nfc related, putting target to 3.1 or later ? 2014-09-15 07:11:22 <[xming]> musl doesn't have uit_t? 2014-09-15 07:11:27 <[xming]> musl doesn't have uid_t? 2014-09-15 07:13:28 [xming]: it does have, you probably don't include something 2014-09-15 07:14:12 musl is less forgiving about such things 2014-09-15 07:14:17 what are you trying to build? 2014-09-15 07:14:21 <[xming]> in glibc it's in sys/type.h I think 2014-09-15 07:14:38 <[xming]> I am trying to build gvpe 2014-09-15 07:16:14 <[xming]> barthalion: thanks a lot , adding #include solved it 2014-09-15 07:16:30 no problem 2014-09-15 07:18:38 vkrishn: feel free to create new ticket for nfc with target 3.1 2014-09-15 07:25:17 http://www.nxp.com/documents/brochure/75017587.pdf - nice 2014-09-15 07:28:07 >>> ERROR: zfs-grsec: Please update _kpkgrel to 1 2014-09-15 07:28:09 ncopa: ↑ 2014-09-15 07:28:18 bah 2014-09-15 07:34:53 <[xming]> cool made my first package 2014-09-15 07:36:44 <[xming]> going for next one, aiccu 2014-09-15 08:08:22 <[xming]> what is the equivelent of __GLIBC__ or __UCLIBC__ for detecting musl? 2014-09-15 08:09:29 nifty, https://delete.im/ 2014-09-15 08:17:51 [xming]: http://wiki.musl-libc.org/wiki/FAQ#Q:_why_is_there_no_MUSL_macro_.3F 2014-09-15 08:18:43 <[xming]> hmm thanks, for quick porting I just test __GLIBC__ instead of _LINUX 2014-09-15 08:19:13 [xming]: that sounds like a good idea 2014-09-15 08:19:37 #if __GLIBC__ #else #endif 2014-09-15 08:19:46 <[xming]> a lot of people don't understand Linux != redhat/ubuntu or even != GNU/glibc 2014-09-15 08:19:52 yup 2014-09-15 08:20:18 <[xming]> ncopa: very nice distro, always wanted to start one since I learnt about musl like 2 years ago 2014-09-15 08:20:25 <[xming]> but never finished it 2014-09-15 08:20:32 much work 2014-09-15 08:20:39 <[xming]> started basing on Getnoo and opkg 2014-09-15 08:20:46 ha 2014-09-15 08:20:57 i starated with alpine linux with gentoo too :) 2014-09-15 08:21:08 <[xming]> I noticed that :D 2014-09-15 08:21:30 <[xming]> still using gentoo, but I want light weight VMs/containers 2014-09-15 08:21:53 alpine is good on that 2014-09-15 08:22:08 specially containers can become very thin 2014-09-15 08:22:16 <[xming]> it's what I had in my mind 2014-09-15 08:22:34 <[xming]> you've done a great job so I don't have to do it 2014-09-15 08:23:34 thanks 2014-09-15 08:24:27 <[xming]> I am porting some pkgs I need to migrate my VMs to LXC + alpine 2014-09-15 08:24:41 which pkgs? 2014-09-15 08:24:52 i can review if you want 2014-09-15 08:24:59 <[xming]> just done gvpe and now aiccu 2014-09-15 08:25:09 ok 2014-09-15 08:25:15 <[xming]> gvpe == L2 VPN 2014-09-15 08:25:21 <[xming]> aiccu == ipve with Siccs 2014-09-15 08:25:29 <[xming]> aiccu == ipv6 with Siccs 2014-09-15 08:25:52 apk add sprunge && sprunge < APKBUILD and i'll have a short look at it 2014-09-15 08:26:00 re musl 2014-09-15 08:26:06 i must say i am pretty impressed with it 2014-09-15 08:26:27 high quality code 2014-09-15 08:26:43 nice team 2014-09-15 08:26:57 they take security seriously 2014-09-15 08:27:12 <[xming]> what is sprunge? 2014-09-15 08:27:24 pastebin like tool 2014-09-15 08:27:49 its a tiny script that wraps curl: http://sprunge.us/ 2014-09-15 08:30:34 shame it's located in the us though:) 2014-09-15 08:31:51 <[xming]> ncopa: gvpe -> http://sprunge.us/VcAg 2014-09-15 08:32:43 patch -p1 -i "$srcdir"/compile-fix.patch || return 1 2014-09-15 08:33:06 in case we move dirs around in future... 2014-09-15 08:33:45 i also use a loop normally, for i in $source; do case $i in *.patch) patch ... 2014-09-15 08:33:54 so its easy to add new patches in future 2014-09-15 08:34:58 looks good 2014-09-15 09:03:20 <[xming]> ncopa: better? http://sprunge.us/TLIb 2014-09-15 09:07:49 yeah 2014-09-15 09:08:26 can you git commit it and: git format-patch --stdout -1 | sprunge 2014-09-15 09:08:41 then i can curl | git am 2014-09-15 09:09:04 <[xming]> ncopa: I am going to tweak it for proper cypher + hashing 2014-09-15 09:09:16 ok 2014-09-15 09:09:29 <[xming]> after that I will 2014-09-15 09:15:39 <[xming]> ncopa: ah I need to add init script too 2014-09-15 09:16:21 <[xming]> thank god it's all openrc :D 2014-09-15 09:18:36 <[xming]> how do I install my own pkg? 2014-09-15 09:24:06 [xming]: I have in /etc/abuild.conf: REPODEST=$HOME/packages/ 2014-09-15 09:24:27 then in my /etc/apk/repositories: /home/ncopa/packages/main 2014-09-15 09:24:33 and testing too 2014-09-15 09:24:43 then its just to apk add it 2014-09-15 09:24:55 btw i also notice in my Xorg.0.log a dbus error: http://sprunge.us/IIDL when i launch dbus-daemon --system it gives a warning about systemd missing 2014-09-15 09:25:04 <[xming]> ncopa: okidoki 2014-09-15 09:25:16 everything works fine, i refuse to use systemd though :) 2014-09-15 09:26:23 Evil_Bob: sounds like it needs dbus yes 2014-09-15 09:26:29 [xming]: you can also do apk add path/to/package.apk 2014-09-15 09:26:51 <[xming]> ah great 2014-09-15 09:27:16 be aware that it will pin it 2014-09-15 09:27:35 so apk upgrade will not automatically upgrade it (unless you apk upgrade -a) 2014-09-15 09:28:28 <[xming]> it doesn't trust my signature :/ 2014-09-15 09:29:22 ah, yes, you need to do abuild-keyring -something 2014-09-15 09:29:24 -a -i? 2014-09-15 09:32:24 <[xming]> did that and copied key to /etc/apk/keys/ 2014-09-15 09:32:48 <[xming]> hmm 2014-09-15 09:32:49 copying key to /etc/apk/keys should do it 2014-09-15 09:33:36 <[xming]> meh PEBCAK 2014-09-15 09:34:37 <[xming]> ok it's in the right dir now, it doesn't say it's untrusted any more but 2014-09-15 09:34:40 <[xming]> ERROR: gvpe-2.25-r0: BAD signature 2014-09-15 09:35:15 <[xming]> hehe, -f won't force it :p 2014-09-15 09:35:26 sounds like it got corrupt? 2014-09-15 09:35:34 apk verify /path/to/file.apk 2014-09-15 09:35:50 tar -ztf /path/to/file.apk 2014-09-15 09:36:14 <[xming]> # apk verify gvpe-2.25-r0.apk gvpe-2.25-r0.apk: 0 - OK 2014-09-15 09:36:35 but apk add wont add it? 2014-09-15 09:36:37 thats weird 2014-09-15 09:36:49 <[xming]> install by path works 2014-09-15 09:37:06 <[xming]> so I will look into stray config/files 2014-09-15 09:38:41 <[xming]> hmm apk is certainly looking for the right file, if I move the gvpe-2.25-r0.apk to gvpe-2.25-r0.apk.bak 2014-09-15 09:38:53 <[xming]> (1/1) Installing gvpe (2.25-r0) ERROR: gvpe-2.25-r0: No such file or directory 2014-09-15 09:39:17 <[xming]> moving it back it says 2014-09-15 09:39:19 <[xming]> ERROR: gvpe-2.25-r0: BAD signature 2014-09-15 09:39:30 <[xming]> but install by path works 2014-09-15 09:39:41 sounds like the apk index has different checksum 2014-09-15 09:39:57 try regenerate the APKINDEX 2014-09-15 09:40:45 in the aports dir: abuild update_abuildrepo_index 2014-09-15 09:42:05 <[xming]> idd that was it, I just deleted the whole dir and rebuild with abuil 2014-09-15 09:56:15 <[xming]> ncopa: gvpe apk -> http://sprunge.us/XfTL 2014-09-15 10:03:46 nice 2014-09-15 10:04:05 i think we have called every single init.d script for *.initd 2014-09-15 10:04:12 so might be nice to be consistent 2014-09-15 10:04:24 mv gvpe.rc gvpe.initd 2014-09-15 10:04:31 but thats just nitpick 2014-09-15 10:06:13 <[xming]> want me to change it? 2014-09-15 10:09:03 <[xming]> http://sprunge.us/LcGa 2014-09-15 10:09:49 <[xming]> oops, just done the git-fu 2014-09-15 10:09:59 <[xming]> but didn't actually update the build file :/ 2014-09-15 10:11:42 <[xming]> this is the one with updated filename http://sprunge.us/Aehe 2014-09-15 10:47:32 <[xming]> what is happening here? http://sprunge.us/HNKA 2014-09-15 10:47:46 <[xming]> I can create a package w/o subpkg 2014-09-15 11:08:10 [xming]: can i see the aiccu apkbuild? 2014-09-15 11:09:17 <[xming]> ncopa: http://sprunge.us/HCEd 2014-09-15 11:10:09 ha 2014-09-15 11:10:16 looks like busybox ash bug 2014-09-15 11:10:41 <[xming]> # install -D -m0755 ${_builddir}/unix-console/aiccu ${pkgdir}/usr/sbin/aiccu# make DESTDIR=${pkgdir} install || return 1 2014-09-15 11:10:52 <[xming]> one of those 2 lines should be uncommented 2014-09-15 11:11:05 <[xming]> but doens't matter. it spews the same error 2014-09-15 11:11:08 problem is the subpackages line 2014-09-15 11:11:23 subpackages="${pkgname}-doc" 2014-09-15 11:11:25 is correct 2014-09-15 11:11:54 i am amused that ${pkgname-doc} didnt throw an error 2014-09-15 11:12:07 <[xming]> nope 2014-09-15 11:12:11 <[xming]> okay got it 2014-09-15 11:12:17 it returned the value of ${pkgname} 2014-09-15 11:12:21 <[xming]> obivious error on my end 2014-09-15 11:12:35 but i think it is an ash bug there too 2014-09-15 11:12:45 ncopa-desktop:~$ foo=hello 2014-09-15 11:12:45 ncopa-desktop:~$ echo ${foo-doc} 2014-09-15 11:12:45 hello 2014-09-15 11:14:03 <[xming]> hehe funny bug :D 2014-09-15 11:14:19 <[xming]> should return "" 2014-09-15 11:14:36 i dont think it is a valid variable name 2014-09-15 11:14:48 you cannot have - in variable names 2014-09-15 11:14:52 so i think it should error 2014-09-15 11:15:00 <[xming]> lolly bash does the same thing 2014-09-15 11:15:06 it does? 2014-09-15 11:15:08 interesting 2014-09-15 11:15:13 and dash? 2014-09-15 11:15:45 does the same 2014-09-15 11:15:54 probably standard feature then 2014-09-15 11:15:59 <[xming]> now where do I have dash 2014-09-15 11:16:03 thx for "I think we should respect polkit-free setups" - /me moved from debian to alpine mainly because of this poettering-stuff and was really annoyed that the login managers of alpine have polkit as dependency :-) 2014-09-15 11:16:36 <[xming]> ncopa: yeap dash does that too 2014-09-15 11:17:06 jomat: i think maybe slim does not? 2014-09-15 11:17:21 [xming]: i think its a posix shell feature 2014-09-15 11:17:31 slim also brings polkit 2014-09-15 11:17:37 oh :-/ 2014-09-15 11:17:41 <[xming]> they all do that with "-" and "\" 2014-09-15 11:17:46 I have this line in my inittab (thx to isaac): ::respawn:/bin/su -c "xinit 2>/dev/null >&2" -l jomat 2014-09-15 11:17:50 <[xming]> really weird 2014-09-15 11:17:58 are there any login manager without polkit dep? 2014-09-15 11:18:02 [xming]: i think its a feature 2014-09-15 11:18:04 and have removed slim 2014-09-15 11:18:21 wait, i'm arriving in frankfurt main station and have to leave :-) 2014-09-15 11:18:54 jomat: please let me know if you find a usable polkit-free loginmanager 2014-09-15 11:19:51 [xming]: http://pubs.opengroup.org/onlinepubs/009604599/utilities/xcu_chap02.html#tag_02_06_02 2014-09-15 11:19:56 there is a table there 2014-09-15 11:20:05 ${parameter-word} 2014-09-15 11:20:05 2014-09-15 11:20:05 substitute parameter 2014-09-15 11:20:05 2014-09-15 11:20:05 substitute null 2014-09-15 11:20:06 2014-09-15 11:20:08 substitute word 2014-09-15 11:20:38 i should have known that... 2014-09-15 11:30:27 <[xming]> really weird feature :p 2014-09-15 11:30:35 nah, manes sense 2014-09-15 11:30:41 i get disconected to freenode 2014-09-15 11:31:54 <[xming]> in sovjiet Russia, freenode disconnects you 2014-09-15 11:35:12 lol 2014-09-15 11:38:26 whatta 2014-09-15 11:38:39 i think i messed up xf86-video-intel 2014-09-15 12:02:16 <[xming]> gvpe apk -> http://sprunge.us/ZYVB 2014-09-15 12:03:18 <[xming]> aiccu apk -> http://sprunge.us/LDKI 2014-09-15 12:03:37 nice 2014-09-15 12:03:43 oh 2014-09-15 12:04:00 it shoudl be introduced in the testing/ dir 2014-09-15 12:04:07 <[xming]> those 2 packages aren't updated often, rather stable ones 2014-09-15 12:04:10 since you added me as maintainer 2014-09-15 12:04:21 <[xming]> so low maintanance :D 2014-09-15 12:04:27 i think i can clean up is you want 2014-09-15 12:04:44 <[xming]> ncopa: sorry, just copied that from other APKBUILD 2014-09-15 12:04:54 thats normal :) 2014-09-15 12:04:57 and you did very good 2014-09-15 12:05:19 do you want me clean it up for you or you want do it yourself? 2014-09-15 12:07:13 <[xming]> ncopa: thanks, been doing my fair share of LFS, gentoo ebuilds, so this very familiar, but better than RPM spec or deb 2014-09-15 12:07:49 <[xming]> ncopa: what kind of clean ups? Go ahead, so I can do a git diff to learn from it 2014-09-15 12:08:15 new packages always goes into testing/ repository 2014-09-15 12:08:47 mostly to check that build does not behave differently on the build server than on your dev env 2014-09-15 12:09:08 then the package built from build server should be tested 2014-09-15 12:09:15 and if its ok, we move it to main 2014-09-15 12:09:29 <[xming]> cool 2014-09-15 12:09:40 testing repo can also be used for early testing (eg test binary before we have written init.d script) 2014-09-15 12:09:48 <[xming]> does very realease has it's main/edge/testing? 2014-09-15 12:10:01 <[xming]> every* 2014-09-15 12:10:13 stable releases has only main 2014-09-15 12:10:17 so we have v3.0/main 2014-09-15 12:10:27 'edge' is the codename for git master 2014-09-15 12:10:38 so edge/main means git master, main subdir 2014-09-15 12:10:40 <[xming]> or testing moves to main, and main get bumped from 3.0.x to 3.x in the future? 2014-09-15 12:10:47 yes 2014-09-15 12:10:57 every 6 months we take git master 2014-09-15 12:11:04 and branch it to a x.y-stable 2014-09-15 12:11:07 and ship iso images 2014-09-15 12:11:16 and maintain that for 2 years 2014-09-15 12:11:25 thats the general idea 2014-09-15 12:11:26 <[xming]> do things get backported to older main? 2014-09-15 12:11:32 <[xming]> ah okay 2014-09-15 12:11:33 on request 2014-09-15 12:11:45 we have do it sometimes 2014-09-15 12:11:48 <[xming]> not really on rolling release 2014-09-15 12:11:52 but try to avoid 2014-09-15 12:12:07 you can consider 'edge' as a rolling release 2014-09-15 12:12:17 i use 'edge' on my desktop 2014-09-15 12:12:42 <[xming]> is edge same as testing or even more bleeding edge? 2014-09-15 12:13:07 <[xming]> I am sticking with Gentoo on my Desktop for now 2014-09-15 12:13:23 edge is git master 2014-09-15 12:13:36 we have main/ and testing/ subdir 2014-09-15 12:13:39 <[xming]> Oh BTW there are some minor issues with the lxc-alpine template 2014-09-15 12:13:58 we dont build testing/ subdir/repo for our stable branches 2014-09-15 12:14:10 <[xming]> are you the maintainer of that too? 2014-09-15 12:14:13 so if you want the package included in v3.1, it needs to move to main 2014-09-15 12:14:15 yes 2014-09-15 12:14:42 and when subitting new features, you should do that on git master (edge) 2014-09-15 12:15:05 commit message for new packages shoudl be: 2014-09-15 12:15:12 testing/: new aport 2014-09-15 12:15:18 2014-09-15 12:15:28 short desc and url 2014-09-15 12:15:46 i normally just copy the $pkgdesc and $url 2014-09-15 12:16:22 it is to show what the package is/does 2014-09-15 12:16:35 and make it easy to find more details if needed (the url) 2014-09-15 12:17:11 <[xming]> so I should run edge/main for my building container 2014-09-15 12:17:18 correct 2014-09-15 12:18:07 <[xming]> and to get pkg from "testing" I have to build them locally 2014-09-15 12:18:21 no, you can test them from edge/testing repo 2014-09-15 12:18:37 <[xming]> binaries are built for testing too? 2014-09-15 12:18:43 the 'edge'/'git master' build server build testing packages 2014-09-15 12:19:09 the build servers for stable releases does not 2014-09-15 12:19:14 <[xming]> ah no testing binaries for "stable" 2014-09-15 12:19:20 correct 2014-09-15 12:19:35 the idea is, we dont ship things that is in "testing" stage 2014-09-15 12:19:41 things that is not 'ready' 2014-09-15 12:20:44 <[xming]> how can I just install some "selected" pkgs from stable/testing? 2014-09-15 12:21:37 <[xming]> imaging my gvpe and aiccu are in stable/testing and I want to add them from the project repo instead of my local repo 2014-09-15 12:21:57 repository pinning: http://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Repository_pinning 2014-09-15 12:24:31 <[xming]> thanks 2014-09-15 12:25:58 i also like to use "" around dir variables 2014-09-15 12:26:06 in case someone installs aports dir in a dir with spaces 2014-09-15 12:26:15 so "$srcdir" or "${srcdir}" 2014-09-15 12:26:37 i know gentoo is consequent on using ${var} instead of $var 2014-09-15 12:26:49 <[xming]> ah correct 2014-09-15 12:27:17 <[xming]> no pptp? 2014-09-15 12:27:21 i tend to use ${var} only where it makes sense 2014-09-15 12:27:45 we have testing/pptpd and testing/pptpclient 2014-09-15 12:27:49 i havent used those 2014-09-15 12:27:59 but might be they dont build 2014-09-15 12:28:07 <[xming]> pptp needs an apkbuild :D 2014-09-15 12:28:22 <[xming]> oh testing 2014-09-15 12:28:50 it means someone has asked for it, someone has built it and pushed it to testing, and never got any feedback if its working or not 2014-09-15 12:29:22 <[xming]> it doesn't build 2014-09-15 12:29:30 oh, re gvpe 2014-09-15 12:29:40 you dont need add openssl and libev to depends 2014-09-15 12:29:51 unless you need run /usr/bin/openssl binary 2014-09-15 12:30:07 abuild will figure out the shared-object depends 2014-09-15 12:30:21 so it is normally only needed to add the -dev package to makedepends 2014-09-15 12:30:29 <[xming]> Oh 2014-09-15 12:30:44 <[xming]> auto dependencies? nice 2014-09-15 12:30:53 yup 2014-09-15 12:31:08 it will add: depends="so:libssl.so.1" or similar 2014-09-15 12:31:15 <[xming]> ACTION gives ncopa thumbs up 2014-09-15 12:31:26 and provides="so:libssl.so.1" to the package that provides it 2014-09-15 12:32:02 <[xming]> I gotta fix that pptpd 2014-09-15 12:32:16 would be good 2014-09-15 12:32:17 thanks 2014-09-15 12:33:07 <[xming]> I need that for my migration 2014-09-15 12:33:51 one big mistage i did with abuild 2014-09-15 12:33:56 misstake* 2014-09-15 12:34:08 was to not run the functions in sh -e 2014-09-15 12:34:22 so it exits on command failure 2014-09-15 12:34:54 so now to be strictly "correct", we need add || return 1 to almost everything 2014-09-15 12:35:33 so i am thinking of making a new build recipy format 2014-09-15 12:36:30 <[xming]> I also like more built-in funcs 2014-09-15 12:36:45 like gentoo does? 2014-09-15 12:36:57 <[xming]> and maybe a a lib loading like gentoo's eclass 2014-09-15 12:37:16 i tried to make the format simple 2014-09-15 12:37:18 <[xming]> yes, it will make APKBUILD shorter and more similar 2014-09-15 12:37:35 so you don't need learn a full api to get started 2014-09-15 12:37:46 but it leads to lots of repetitive code 2014-09-15 12:37:57 <[xming]> that's nice too, ebuilds were like this in 2002 :D 2014-09-15 12:38:32 <[xming]> but as no. of pkgs grow, the maintananbility of the whole repo is more important 2014-09-15 12:38:37 yes 2014-09-15 12:38:41 that is a good point 2014-09-15 12:39:17 <[xming]> it's still possible to write ebuilds the way it was 2014-09-15 12:39:40 <[xming]> those just won't get commited 2014-09-15 12:39:54 i think if i change the format it will need break backwards compat 2014-09-15 12:40:21 gentoo did that pretty good actually 2014-09-15 12:40:30 that they didnt break the backwards compat 2014-09-15 12:40:33 <[xming]> better to have that now than 5 years later when Alpine is popular :D 2014-09-15 12:40:51 it will be difficult enough now... 2014-09-15 12:41:43 <[xming]> have a versioning filed in APKBUILD 2014-09-15 12:42:15 <[xming]> when it's the old verion, invoke the old one 2014-09-15 12:42:16 apiversion 2014-09-15 12:42:20 or so 2014-09-15 12:42:21 but 2014-09-15 12:42:37 <[xming]> only use new version with newer APKBUILD 2014-09-15 12:42:48 i'd like to get rid of filtering it through shell when parsing the entire tree 2014-09-15 12:43:09 so no more while/for loops in global scope 2014-09-15 12:43:10 <[xming]> so you can phase out the old APKBUILD instead of a cut off 2014-09-15 12:43:51 i'd like to parse the file with lua 2014-09-15 12:44:17 so i think it would become more similar redhats .spec file with sections 2014-09-15 12:44:41 maybe i should post an email to the devel mailing list about my thoughts there 2014-09-15 12:45:55 <[xming]> Gentoo have eclass for autotools, so no need to do DESTDIR=... make ... 2014-09-15 12:46:20 <[xming]> autotools is used by probably 60% of pkgs and other 40% is almost all cmake 2014-09-15 12:46:57 you need learn an api... 2014-09-15 12:46:59 but yes 2014-09-15 12:47:10 <[xming]> I am just saying it's handy, and easy to maintain your repo of 60K+ pkgs :D 2014-09-15 12:47:11 it makes maintenance easier 2014-09-15 12:47:13 yes 2014-09-15 12:47:15 definitively 2014-09-15 12:47:27 i maintain 1000+ so i know what you talk about :) 2014-09-15 12:48:10 btw, the gvpe example config and .initd and .confd has whitespace damages 2014-09-15 12:48:17 lines that ends with whitespace 2014-09-15 12:48:32 and space/tab mix 2014-09-15 12:48:39 + -- -linfo -L ${NODENAME} 2014-09-15 12:49:12 here is modified commit: http://sprunge.us/CXYZ 2014-09-15 12:50:58 <[xming]> you don't like ending \n ? 2014-09-15 12:51:12 <[xming]> or better \n\n ? 2014-09-15 12:51:14 i dont like space before \n 2014-09-15 12:51:34 space/tab before \n 2014-09-15 12:51:55 <[xming]> I dont have them in my files 2014-09-15 12:52:07 you have, you just gannot see them ;) 2014-09-15 12:52:14 cannot* 2014-09-15 12:52:36 sed -i -E -e 's/ +$//' gvpe.conf.example gvpe.confd gvpe.initd if-up && git diff 2014-09-15 12:54:01 i'm pushing testing/gvpe now 2014-09-15 12:54:19 thanks! 2014-09-15 13:01:36 <[xming]> thanks, what needs to be done so that it will land in main? 2014-09-15 13:07:36 test it and if it works, drop someone here a line 2014-09-15 13:22:04 did somebody try to build emacs ? 2014-09-15 13:26:15 <[xming]> ncopa: fixed pptpd and bumped to 1.40 -> http://sprunge.us/KceF 2014-09-15 13:28:24 <[xming]> building emacs :O 2014-09-15 13:29:13 doesnt work muslproblems 2014-09-15 13:29:29 if you manage im your fan 2014-09-15 13:30:20 <[xming]> what doesn't work? Building emacs? 2014-09-15 13:30:23 yeah 2014-09-15 13:30:34 emacs does nasty things 2014-09-15 13:30:41 like implement its own malloc... 2014-09-15 13:30:45 <[xming]> ACTION shuts up about emacs 2014-09-15 13:30:52 which is not supported by musl 2014-09-15 13:31:00 same problem with valgrind 2014-09-15 13:31:26 i think valgrind is the thing I miss most with musl libc 2014-09-15 13:32:31 <[xming]> in case of valgrind it's understanable, but for emacs, wht its own mallo() :/ 2014-09-15 13:32:51 <[xming]> must be the kitchin-sink syndrome 2014-09-15 13:36:18 [xming]: thanks for pptpd fix 2014-09-15 13:36:25 <[xming]> welcome 2014-09-15 13:36:30 i push with a modified commti message 2014-09-15 13:36:46 we use: repo/pkgname: msg 2014-09-15 13:36:59 prefix with main/pkg: or testing/pkg: 2014-09-15 13:37:17 git log --format=oneline 2014-09-15 13:38:04 <[xming]> ah yes you told me that, and I just noticed that my commit msg goes to the main git repo 2014-09-15 13:40:00 <[xming]> is algitbot the auto build bot? 2014-09-15 13:40:32 yeah 2014-09-15 13:40:33 will 2014-09-15 13:40:34 well 2014-09-15 13:40:50 git push hook emits mqtt message 2014-09-15 13:41:09 algitbot subscribes to git push mqtt messages 2014-09-15 13:41:15 so does the build servers 2014-09-15 13:42:33 and redmine server (which auto resolves on git commit messages) 2014-09-15 13:43:16 <[xming]> when the bots says uploaded, does that mean the build was successful and the binary is uploaded? 2014-09-15 13:43:22 correct 2014-09-15 13:43:38 to nl.alpinelinux.org which is the mirror master 2014-09-15 13:43:45 aka rsync.alpinelinux.org 2014-09-15 13:43:58 <[xming]> so it compiles on x86, amd64 and arm now? 2014-09-15 13:44:09 i think it is done already 2014-09-15 13:44:20 <[xming]> yes it did 2014-09-15 13:44:42 <[xming]> just meant if my fix worked on those 3 platforms 2014-09-15 13:45:00 i think it did 2014-09-15 13:46:10 <[xming]> cool 2014-09-15 13:50:39 [xming]: do you think you could clean up the aiccu patch? 2014-09-15 13:50:59 <[xming]> ncopa: an note about the repo structure, some fs still have 16K/32K inodes per dir 2014-09-15 13:51:23 and resend, if i dont apply it today, then please git send-email to alpine-devel@lists.alpinelinux.org 2014-09-15 13:51:43 so we will hit a limit when reaching 16K aports? 2014-09-15 13:53:12 <[xming]> yes 2014-09-15 13:53:21 need think of that too 2014-09-15 13:53:28 <[xming]> on some old FS 2014-09-15 13:53:44 FAT will not work anyways 2014-09-15 13:53:49 since we use symlinks 2014-09-15 13:54:00 but it might be nice to support ext2 2014-09-15 13:54:06 <[xming]> ncopa: what kind of cleanups do you have in mind? whitespace? 2014-09-15 13:54:19 commit message mostly 2014-09-15 13:54:26 maybe "$srcdir" 2014-09-15 13:54:32 or "${srcdir}" 2014-09-15 13:54:37 in case spaces in dirname 2014-09-15 13:55:52 <[xming]> ah okay 2014-09-15 13:55:52 maybe also exit on error if some of the extra installed files, init.d or example config is missing 2014-09-15 13:55:59 <[xming]> it seems that even ext3 has 32K limit 2014-09-15 13:56:29 32K is likely not a problem 2014-09-15 13:57:08 but there are other reasons we might want rethink dir structure too 2014-09-15 13:57:50 i have been thinking in having 'main' as a more long-time-support dir 2014-09-15 13:58:00 and have a short-time support dir for some things 2014-09-15 13:58:33 there are some packages that we cannot support as long as we would like, due to upstream 2014-09-15 13:58:46 <[xming]> most FS goes really slow with large number of inodes 2014-09-15 13:58:52 firefox and qemu are good examples 2014-09-15 13:58:54 also good reason 2014-09-15 13:59:01 oh i need to go. brb 2014-09-15 13:59:32 <[xming]> 32K isn't that much :p 2014-09-15 13:59:38 <[xming]> ciao 2014-09-15 14:12:19 <[xming]> ncopa: hopefully I've done it right this time -> http://sprunge.us/HJdh 2014-09-15 15:12:17 >>> aiccu: Checking sha512sums... 2014-09-15 15:12:17 aiccu_20070115.tar.gz: OK 2014-09-15 15:12:17 musl_does_not_have_res_ninit.patch: OK 2014-09-15 15:12:17 fix_linking.patch: OK 2014-09-15 15:12:17 aiccu.initd: FAILED 2014-09-15 15:12:18 sha512sum: WARNING: 1 of 1 computed checksums did NOT match 2014-09-15 15:12:36 and it need to be moved from main to testing 2014-09-15 15:12:41 need to go home now 2014-09-15 15:12:42 see u 2014-09-15 15:25:18 <[xming]> sigh so manyth attempt :p 2014-09-15 15:25:24 <[xming]> ncopa: http://sprunge.us/VPEc 2014-09-15 16:27:43 [xming]: pushed, next time try to wrap long lines. 2014-09-15 16:39:02 <[xming]> clandmeter: ty 2014-09-15 17:11:39 question, how can I check the contents of a package/port using apk? 2014-09-15 17:34:15 sorry, that question should've been sent to #alpine-linux 2014-09-15 18:48:12 as a reference, ff 29.0 is kinda outdated:/ 2014-09-15 19:15:43 Amnesia: i think testing has 31 2014-09-15 19:16:05 yeah, I'm switching from stable to edge aswp 2014-09-15 19:16:07 yea i use edge, its version 31 2014-09-15 19:16:14 err edge 2014-09-15 19:16:15 edge == testing? 2014-09-15 19:16:18 yeah 2014-09-15 19:16:21 that's what i meant 2014-09-15 19:17:24 hm, lets see whether I'm still able to boot:-) 2014-09-15 19:17:25 we should probably think about shipping firefox similiarly to Ubuntu 2014-09-15 19:17:39 which pushes the newest firefox to all supported releases 2014-09-15 19:18:08 ncopa: ↑ 2014-09-15 19:19:02 w00p, success 2014-09-15 19:19:53 <[xming]> hmm can that be called firefox? 2014-09-15 19:20:49 <[xming]> AFAIK mozilla's policy is that only their builds can be called firefox (and use the logo/icon) 2014-09-15 19:21:05 <[xming]> hence Debian has iceweasel 2014-09-15 19:22:14 Debian has iceweasel because they're adding their own patches 2014-09-15 19:22:46 as long as your sources are vanilla, there is no reason not to have Mozilla's branding 2014-09-15 19:24:31 <[xming]> alpine patches firefox too 2014-09-15 19:24:58 well, then we shouldn't be using firefox name 2014-09-15 19:25:24 <[xming]> it's just a patch against search engine 2014-09-15 19:25:46 <[xming]> but I don't know how strict mozilla at enforcing this 2014-09-15 19:26:07 I've no idea 2014-09-15 19:26:59 I guess it's not that important to call that a violation 2014-09-15 19:28:55 <[xming]> and I am not sure if it's just patches, IIRC it's really their build (with bundled libs, specific version of gcc, ...) 2014-09-15 19:29:54 <[xming]> Gentoo doesn't patch firefox and yet they warn it's illegal to distribute self compiled versions of firefox 2014-09-15 19:30:14 <[xming]> if one wants to distribute that, gentoo turns on the aurora branding 2014-09-16 06:15:50 barthalion: iirc, ubuntu does not ship firefox with shared libs 2014-09-16 06:16:32 they have as many static libs as possible 2014-09-16 06:16:37 that is why they can do so 2014-09-16 06:16:42 but the binary becomes bigger 2014-09-16 06:17:23 i would really prefer use shared, systems libs 2014-09-16 06:17:53 but newer versions of ff sometimes requires newer, non-abi-compat versions of the libs too 2014-09-16 06:18:07 so, i think we have the following options 2014-09-16 06:18:20 1) link libs statically, similar to what ubuntu does 2014-09-16 06:18:33 2) try to stick to ff ESR for stable 2014-09-16 06:18:43 the current plan is to stick to ESR 2014-09-16 06:39:04 <[xming]> morning all 2014-09-16 06:39:22 <[xming]> I see updated pptpd and new aiccu in http://nl.alpinelinux.org/alpine/edge/testing/x86_64/ 2014-09-16 06:39:29 <[xming]> but I can't find gvpe 2014-09-16 06:41:51 humm... http://git.alpinelinux.org/cgit/aports/commit/?id=60bf174308036b0d6e9dec0aaefa9fc07591fdae 2014-09-16 06:42:38 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64/testing/gvpe/ 2014-09-16 06:43:58 checking for inflate in -lz... yes 2014-09-16 06:44:04 i think we need add zlib-dev to makedepends 2014-09-16 06:44:10 checking for gmp.h... no 2014-09-16 06:44:10 configure: error: gmp.h not found, required for --enable-dns 2014-09-16 06:44:16 and we need add gmp-dev too 2014-09-16 06:45:53 <[xming]> thanks, although I can't remember that I installed them, probably pulled when I installed other things 2014-09-16 06:49:45 [xming]: yeah, same here 2014-09-16 06:49:55 i probably had them, so i didnt catch it when i teted 2014-09-16 06:50:00 tested* 2014-09-16 06:50:38 thats one of reasons we have the testing repository, so we dont end up pushing things to 'main' that does not even build... 2014-09-16 06:50:46 <[xming]> grrr, my squid is caching that build log :/ 2014-09-16 06:51:20 <[xming]> ACTION blames chromium 2014-09-16 06:51:31 i suppose i can tell lighttpd to not cache those 2014-09-16 06:51:44 or reduce ttl 2014-09-16 06:51:49 dunno how though 2014-09-16 06:52:17 <[xming]> it's chromium and not my squid 2014-09-16 07:11:02 <[xming]> what is the alpine way configurating dialup (PPPoE)? 2014-09-16 07:11:32 <[xming]> or I just put scripts in /etc/local.d? 2014-09-16 07:18:04 you can use /etc/network/interfaces for it 2014-09-16 07:18:33 auto ppp0 2014-09-16 07:18:33 iface ppp0 inet ppp 2014-09-16 07:18:33 pre-up ip link set dev eth0 up 2014-09-16 07:18:33 provider telenor 2014-09-16 07:19:02 /etc/ppp/peers/telenor 2014-09-16 07:19:26 the 'provider myprovider' needs a file in /etc/ppp/peers/ 2014-09-16 07:20:23 in my /etc/ppp/peers/telenor (telenor is the name of one of my ISPs) I have: 2014-09-16 07:20:25 # rp-pppoe plug-in makes PPPoE connection so rp-pppoe package is not needed 2014-09-16 07:20:26 # Possibly, you may need to change interface according your configuration 2014-09-16 07:20:26 plugin rp-pppoe.so eth0 2014-09-16 07:23:18 <[xming]> ah debian way thanks 2014-09-16 07:54:18 <[xming]> I have just tested aiccu, it works as it should with iproute2 installed 2014-09-16 07:54:51 <[xming]> with bb's ip the default route doesn't get added 2014-09-16 07:55:38 <[xming]> I think makedepends="gnutls-dev iproute2" should be 2014-09-16 07:55:45 <[xming]> makedepends="gnutls-dev" 2014-09-16 07:55:53 <[xming]> depends="iproute2" 2014-09-16 07:55:57 yes 2014-09-16 07:56:00 sounds correct 2014-09-16 07:56:46 http://sprunge.us/EMSC 2014-09-16 08:00:58 <[xming]> aiccu tested and works for me (tm), same setup as my current sixxs setup 2014-09-16 08:01:16 <[xming]> # ping6 google.com PING google.com (2a00:1450:400c:c03::8b): 56 data bytes 2014-09-16 08:01:23 <[xming]> 64 bytes from 2a00:1450:400c:c03::8b: seq=0 ttl=55 time=46.466 ms 2014-09-16 08:01:32 ok 2014-09-16 08:01:36 we can move it to main then 2014-09-16 08:01:57 <[xming]> init script stop/start work as it should 2014-09-16 08:03:13 <[xming]> ncopa: thanks a lot 2014-09-16 08:03:38 no, thank *you* for contributing :) 2014-09-16 08:04:36 <[xming]> welcome, I need those pkgs to achieve what I want :D So better to make apkgs of them 2014-09-16 08:04:59 thats the way open source is supposed to work 2014-09-16 08:05:49 <[xming]> idd, so glad to have a decent distro w/o udev/pulse/policykit/systemd/... polutions 2014-09-16 08:06:31 <[xming]> might not be the best desktop experience for newbies, but this is certainly the best way for VM/containers 2014-09-16 08:07:39 <[xming]> ncopa: ah I thought about something else, if you put everything in one directory, beside hitting 16k/32k some day, the is also risk of name collision 2014-09-16 08:12:50 yea, we need to reorganize dirs 2014-09-16 08:12:56 dont know how though 2014-09-16 08:13:02 and now i need coffe... 2014-09-16 08:16:12 there aren't any plans to implement systemd right!?:P 2014-09-16 08:20:03 <[xming]> I don't mind systemd, as long as it's a choice, but by the looks of it unless a distro has chosen to avoid systemd, it won't be a choice any more 2014-09-16 08:20:23 <[xming]> Gentoo is give me that choice, but I wonder for how long 2014-09-16 08:20:37 <[xming]> s/is/still 2014-09-16 08:21:44 <[xming]> udev and pulse were a PITA, but I learnt to deal with them :/ 2014-09-16 08:22:22 :-) 2014-09-16 08:22:51 hm, has one of you folks succeeded in getting X to work succesfully without udev? 2014-09-16 08:25:51 Amnesia: i think it is possible but you will likely need configure it manually then 2014-09-16 08:26:06 and you will lose keyboard/mouse hotplug support 2014-09-16 08:26:06 <[xming]> it's a pity that most of the xorg people are going down that route 2014-09-16 08:26:25 ah 'k 2014-09-16 08:26:42 that explains why the keyboard/mouse stops working when I ditch udev^^ 2014-09-16 08:26:51 it would have been okish, unless udev people would try make it hard to run udev without systemd 2014-09-16 08:27:04 <[xming]> like the good old days 2014-09-16 08:27:26 Amnesia: it should work if you manually configure keyb mouse in xorg.conf 2014-09-16 08:27:48 <[xming]> they stopped maintaining udev as seperate project, only support the systemd's udev 2014-09-16 08:27:56 i know 2014-09-16 08:28:00 thats the sad part 2014-09-16 08:28:17 and i think that was strategic move 2014-09-16 08:28:23 definitely 2014-09-16 08:28:27 get most people on the udev train 2014-09-16 08:28:44 when the cost for ditching udev support is high enough for most people 2014-09-16 08:28:45 <[xming]> well some of us don't like the Mac/Windows way of handling things autogimatically 2014-09-16 08:28:59 have them over on systemd 2014-09-16 08:29:14 <[xming]> personally I like static /dev more than udev 2014-09-16 08:29:41 its kinda nice to not need manually write an xorg.conf file... 2014-09-16 08:30:02 <[xming]> devtmpfs/mdev is the best middle ground 2014-09-16 08:30:25 mdev has its own set of issues 2014-09-16 08:30:45 i have been thinking of writing a mdevd 2014-09-16 08:30:57 that reads netlink socket 2014-09-16 08:30:58 <[xming]> everything has issues, but I rather deal with all of them then udev 2014-09-16 08:31:02 uses mdev.conf 2014-09-16 08:31:32 and i think it would be nice to support a subset of libudev too 2014-09-16 08:31:38 so we could run things like xorg 2014-09-16 08:31:48 <[xming]> no matter what I do, udev still messes with my 4 port ehternet NICs just found a new way to do it hopefully it stays that way 2014-09-16 08:32:32 <[xming]> with static devs they only change naming when PCI orders changes 2014-09-16 08:32:46 <[xming]> with udev they have their own minds 2014-09-16 08:33:31 i think that was one goals with udev 2014-09-16 08:34:11 make network interface names named to what ever you want 2014-09-16 08:34:18 and make it persistent 2014-09-16 08:36:38 ncopa: hello, i'm trying to build emacs (without x), it sad that stderr is read-only, i checked on other distro stdio.h, and stderr is struct, what i should do to compile successfully? 2014-09-16 08:37:16 ? 2014-09-16 08:37:25 stderr is readonly? 2014-09-16 08:37:41 sounds like emacs is trying to do things with glibc internals 2014-09-16 08:37:56 ncopa: yes, in /usr/include/stdio.h stderr has type *const 2014-09-16 08:38:06 i know they try to do nasty things with malloc 2014-09-16 08:38:18 or provide their own malloc 2014-09-16 08:38:24 which does not work on musl 2014-09-16 08:38:57 i suppose the best way would be to submit bug to upstream emacs 2014-09-16 08:39:04 :/ 2014-09-16 08:39:08 "does not build with musl libc" 2014-09-16 08:39:12 with error message 2014-09-16 08:39:21 i dont think it will be easy to fix it 2014-09-16 08:39:27 <[xming]> ncopa: indeed, but it's totally broken here :( 2014-09-16 08:39:42 very sad 2014-09-16 08:40:15 k0r10n: i have seen better code quality than what emacs provides... :-/ 2014-09-16 08:40:20 <[xming]> time to migrate to vi? :p 2014-09-16 08:40:26 <[xming]> ACTION hides 2014-09-16 08:40:28 vim works great ;) 2014-09-16 08:40:34 [xming]: i'm vim user :) 2014-09-16 08:40:56 [xming]: i just need emacs to compile lisp code 2014-09-16 08:41:12 <[xming]> vee-eye 2014-09-16 08:41:55 http://stackoverflow.com/questions/913671/are-there-lisp-native-code-compilers 2014-09-16 08:43:00 ty ncopa, i will look at it 2014-09-16 08:43:12 points to this: http://www.thefreecountry.com/compilers/commonlisp.shtml 2014-09-16 08:48:15 <[xming]> gvpe tested and working with my existing config 2014-09-16 08:48:42 <[xming]> /usr/sbin/gvpe[310]: games(rawip/95.170.xxx.yyy:0): connection established (direct), protocol version 0.1 2014-09-16 08:49:01 <[xming]> ping 192.168.253.1 2014-09-16 08:49:13 <[xming]> 64 bytes from 192.168.253.1: seq=0 ttl=64 time=71.136 ms 2014-09-16 09:34:21 <[xming]> why is masq called snat in awall? what if someone wants SNAT (without PAT)? 2014-09-16 10:35:27 <[xming]> pptpd-1.4.0 tested and works with my current config 2014-09-16 10:35:50 <[xming]> ncopa: both gvpe and pptpd are tested and working 2014-09-16 11:21:34 moved pptpd 2014-09-16 14:24:19 \o/ they accpeted the nfs patches 2014-09-16 19:15:34 ncopa: yay 2014-09-17 06:26:02 xulrunner fails to build on x86: http://bld1.alpinelinux.org/buildlogs/build-edge-x86/testing/xulrunner/xulrunner-32.0.1-r0.log 2014-09-17 06:26:12 xpshell seems to segfault 2014-09-17 06:36:25 <[xming]> morning 2014-09-17 06:37:06 <[xming]> how do I port forwarding to different port using awall? 2014-09-17 06:47:27 [xming]: http://wiki.alpinelinux.org/wiki/How-To_Alpine_Wall#Port-Forwarding 2014-09-17 06:47:38 hm 2014-09-17 06:47:51 does not tell how to do it to different port 2014-09-17 06:48:38 http://wiki.alpinelinux.org/wiki/Alpine_Wall_User%27s_Guide#NAT_Rules 2014-09-17 07:17:46 <[xming]> ncopa to the rescue :D thanks 2014-09-17 07:18:29 actually, it was rnalrd 2014-09-17 07:18:36 [xming] http://wiki.alpinelinux.org/w/index.php?title=How-To_Alpine_Wall&diff=10151&oldid=8706 2014-09-17 07:18:37 <[xming]> only thing that I don't like about json is that I can't put comments in it :/ 2014-09-17 07:18:46 [xming]: yeah, its a pain... 2014-09-17 07:19:31 i acutally just thought if it might be worth to add hack for that, ala sed '/^#/d' 2014-09-17 07:19:55 simply filter out lines that starts with # 2014-09-17 07:20:01 before processing the json 2014-09-17 07:20:32 but then it would not be valid json 2014-09-17 07:21:23 <[xming]> I don't like shorewall splitting things in dozens of different files which I could never find what I want w/o using grep 2014-09-17 07:21:39 heh, yes 2014-09-17 07:21:39 <[xming]> awal is much saner, I can organise the file the way I want 2014-09-17 07:21:53 i have mixed feeling about the json though 2014-09-17 07:21:59 <[xming]> but I miss comment :/ 2014-09-17 07:22:08 even if it was me who pushed for json back at that time 2014-09-17 07:22:23 <[xming]> json is hell lot better than xml :D 2014-09-17 07:22:31 i think yaml would have been nicer 2014-09-17 07:22:49 xml was never considered 2014-09-17 07:22:57 yes, i miss comments 2014-09-17 07:22:59 <[xming]> yaml is nice too 2014-09-17 07:23:12 that was one of the biggest worries when we picked it 2014-09-17 07:23:34 i think there is support for "description": "my comment" 2014-09-17 07:23:52 <[xming]> rnalrd: thanks 2014-09-17 07:24:53 [xming]: awall also lets you group a set of rules into its own .json file 2014-09-17 07:24:57 <[xming]> awall encourages to put as less as possible in a file, as one can toggle each file on/off 2014-09-17 07:25:04 then you can awall enable mygroup 2014-09-17 07:25:10 yes 2014-09-17 07:25:30 <[xming]> put having too much files is also a PITA 2014-09-17 07:25:47 you dont need to use that feature ofcourse ;) 2014-09-17 07:26:19 but its nice as it lets you reuse a set of common rules between many hosts 2014-09-17 07:26:35 lets say you want enable ssh as standard 2014-09-17 07:26:42 on many hosts 2014-09-17 07:26:46 <[xming]> the pain is that I cannot comment out a single rule/filter unless it's in a seperate file 2014-09-17 07:27:07 ah, that is true 2014-09-17 07:27:25 <[xming]> wc -l /etc/shorewall/rules 2014-09-17 07:27:32 <[xming]> 166 /etc/shorewall/rules 2014-09-17 07:27:37 i wonder if it might be worth add a pre filter 2014-09-17 07:27:48 that filters out comments before processing json 2014-09-17 07:30:11 <[xming]> if that's implemented I can foresee myself swearing at the ending commas :p 2014-09-17 07:35:28 so other option i have been thinking of is 2014-09-17 07:35:34 having .yaml in that dir 2014-09-17 07:35:44 and then a makefiel that converts them to json 2014-09-17 07:35:59 something like: http://sprunge.us/PMLN 2014-09-17 07:37:23 could have it run awall translate --verify to verify that the rules are ok 2014-09-17 07:37:34 actually 2014-09-17 07:37:44 it would not be that hard to add yaml support to awall 2014-09-17 07:37:49 optional yaml support 2014-09-17 08:13:02 <[xming]> that sounds great 2014-09-17 08:47:18 <[xming]> why does php depends on php-cgi? I have php-fpm and the moment I installed php-pdo readline/ncurses/php-cgi get pulled :( 2014-09-17 08:48:38 <[xming]> is it possible to make an OR depend? 2014-09-17 08:48:52 <[xming]> depends="$pkgname-cgi | $pkgname-fpm" 2014-09-17 09:11:20 [xming]: i think it is due to compat/history reasons 2014-09-17 09:11:30 in old days there was only a php package 2014-09-17 09:12:16 and we did not want break things for people who had 'apk add php' and used cgi 2014-09-17 09:12:19 and then did apk upgrade 2014-09-17 09:12:30 i suppose we should have fixed that for the v3.0 release 2014-09-17 09:12:42 i suppose we can clean it up for v3.1 2014-09-17 13:24:12 ok, seems like firefox is broken on x86 2014-09-17 13:24:15 does not build 2014-09-17 13:24:36 firefox-31.0 seems to work 2014-09-17 14:18:08 curl 7.38 has some new features, would be nice to have a look at it and if possible enable them 2014-09-17 14:51:23 read somewhere markdown getting standardized 2014-09-17 14:53:18 cool 2014-09-17 14:54:28 I installed a fresh chroot with "sudo apk -X ${mirror}/v3.0/main -U --allow-untrusted --root ${chroot_dir} --initdb add alpine-base", added alpine-sdk, and abuild fails with "/usr/bin/abuild: .: line 1969: : not found" ... anyone any idea? 2014-09-17 14:56:08 No grsec-violations in dmesg 2014-09-17 14:56:22 jomat: can you give a link to the script you use? 2014-09-17 15:00:19 Evil_Bob: https://www.jmt.gr/~jomat/xdm.tar.bz2 2014-09-17 15:02:39 jomat: maybe something wrong in your /etc/abuild.conf? 2014-09-17 15:04:30 I copied it from a working build environment... it looks ok 2014-09-17 15:04:32 jomat: i dont have the deps specified in that APKBUILD, but it works for me 2014-09-17 15:05:07 Evil_Bob: You need some testing package for it 2014-09-17 15:07:26 That's my abuild.conf: https://0.jmt.gr/?e30ef40f58fef512#EjWSlQpCcYTM6Uu4SE5++0/ZX+miWeSdehgHy47RsLQ= 2014-09-17 15:08:58 can you post it on sprunge.us or so? 2014-09-17 15:11:14 wget --no-check-certificate -qO- https://www.jmt.gr/~jomat/abuild.conf 2014-09-17 15:11:49 Wait 2014-09-17 15:11:54 That looks wrong 2014-09-17 15:14:24 Evil_Bob: try again, now it's the right file 2014-09-17 15:16:25 your abuild.conf looks fine 2014-09-17 15:35:46 Oh ok... readlink -f ./APKBUILD doesn't work in the chroot 2014-09-17 15:37:14 Because i need /proc 2014-09-17 15:38:06 Now it works :-D 2014-09-17 15:41:57 nice :) 2014-09-18 06:27:50 jomat: i tend to prefer lxc over chroot 2014-09-18 06:29:37 algitbot:build 2.7-stable 2014-09-18 06:35:21 ncopa, how do I force rebuilding of 2.7-stable? 2014-09-18 06:39:02 mosquitto_pub -h msg.alpinelinux.org -t git/aports/2.7-stable -m retry 2014-09-18 06:39:05 or so 2014-09-18 06:54:19 tnx 2014-09-18 06:54:25 all good now 2014-09-18 07:03:33 it was procmail apkbuild that did weird things 2014-09-18 07:03:42 like: yes n | make ... 2014-09-18 07:03:57 'yes' never exited 2014-09-18 07:10:13 question, does stuff often break in testing? 2014-09-18 07:20:01 yes 2014-09-18 07:20:16 as in, does not build anymore 2014-09-18 07:20:30 hm ok 2014-09-18 07:20:38 the idea of testing was 2014-09-18 07:20:46 a developer build soemthing locally 2014-09-18 07:20:46 to test:P? 2014-09-18 07:20:52 test it locally 2014-09-18 07:20:56 and things just work 2014-09-18 07:21:08 but will it work if it is built on the build server? 2014-09-18 07:21:20 because local dev might have other packages installed at build time 2014-09-18 07:21:27 so he pushes the package to testing 2014-09-18 07:21:36 and can test that the autobuilt package also works 2014-09-18 07:21:43 or have someone else to test it 2014-09-18 07:22:01 when confirmed that it actually works, it is supposed to be moevd to main 2014-09-18 07:22:40 so, for example, someone asks for package foo 2014-09-18 07:22:51 a developer builds it and pushes it to testing 2014-09-18 07:23:04 okay, that actually makes sense:) 2014-09-18 07:23:13 and instad of developer verify that it actually works, he lets other do that 2014-09-18 07:23:22 but as you see 2014-09-18 07:23:24 testing is 'huge' 2014-09-18 07:23:25 dafuq.. when I execute apk upgrade wayland gets installed, while it doesn't have any rdepends o0 2014-09-18 07:23:54 so people tend to foget to give that feedback that it can be moved to main 2014-09-18 08:00:50 ncopa: Hello, i tested partimage and it works fine 2014-09-18 08:11:16 <[xming]> ncopa: I saw an other approach for commenting in json, keys starting with '#' are ignored 2014-09-18 08:12:01 the problem is that there is no standard for it really 2014-09-18 08:13:51 k0r10n: thanks! 2014-09-18 08:57:59 ncopa: i've got cannot map memory to address while compiling clisp - could you please give me advice? 2014-09-18 09:03:03 does dmesg show that grsec is involved? 2014-09-18 09:03:09 does clisp use JIT or similar? 2014-09-18 09:03:23 you might need paxctl -m path/to/binary 2014-09-18 09:06:05 segfault in dmesg... 2014-09-18 09:07:36 ncopa: ok ty, will look later 2014-09-18 12:57:11 <[xming]> any particular reason why mtr has a special patch to disable DNS? 2014-09-18 12:57:45 <[xming]> DNS lookup can be disabled at runtime with -n 2014-09-18 12:58:16 <[xming]> I would like to version bump mtr to 0.85, wonder if that should be kept 2014-09-18 13:03:34 ncopa: about clisp, i've got denied resource overstep in dmesg when trying to compile 2014-09-18 13:08:04 [xming]: i think they use some res_ functions that is not implemented in musl or similar 2014-09-18 13:08:09 i looked at it some time ago 2014-09-18 13:08:13 was not really trivial to fix 2014-09-18 13:08:29 i had hoped for use of c-ares or similar 2014-09-18 13:09:12 some history: https://bugs.launchpad.net/mtr/+bug/1021683 2014-09-18 13:32:41 ncopa: looks like the git tarball trigger is broken for acf-freeradius3 repo 2014-09-18 13:32:44 can you fix it? 2014-09-18 13:32:49 (please) 2014-09-18 13:35:56 its just to create the /archive/acf-freeradius3 dir 2014-09-18 13:36:04 i just did it 2014-09-18 13:37:03 should I manually run the command to create the tarball? 2014-09-18 13:39:25 i did it 2014-09-18 13:39:31 ok, thank you 2014-09-18 13:41:51 i think i should reorganize the hooks 2014-09-18 19:38:27 hi@all 2014-09-18 19:38:40 question: in which package is "chsh" includet? 2014-09-18 20:42:24 n8@all 2014-09-19 08:42:34 ncopa: hello, i tested LXDE, it's working great 2014-09-19 08:44:19 nice 2014-09-19 08:44:38 can you give me a list of the packages you want move to main? 2014-09-19 08:44:48 and verify that all the deps are included 2014-09-19 08:45:00 we cannot have packages in main that depends on packages in testing 2014-09-19 08:49:08 lxappearance, lxappearance-obconf, lxde-common, lxde-icon-theme, lxmenu-data, lxpanel 2014-09-19 08:49:57 no testing dependencies 2014-09-19 10:03:27 <[xming]> new aport http://sprunge.us/dTiF 2014-09-19 10:14:33 [xming]: make sure you have a clean buildenv 2014-09-19 10:18:00 <[xming]> hmm I have a special build env, maybe I installed some -dev manually and forgot to remove them 2014-09-19 10:18:52 check /etc/apk/world 2014-09-19 10:19:05 atleast thats how i do it 2014-09-19 10:19:28 apk info can be cluthered with deps 2014-09-19 10:22:10 <[xming]> remove almost all -dev except musl and libc 2014-09-19 10:23:17 i think ive only got vim install on my buildenv 2014-09-19 10:23:20 and maybe strace 2014-09-19 10:23:46 and of course alpine-sdk 2014-09-19 10:24:07 <[xming]> ah :D 2014-09-19 10:25:25 i need to shoot my provider, by providing me such a shitty cisco modem... it keeps cleaning my nat table after a while. 2014-09-19 10:26:10 <[xming]> I don't trust anything else beside linux/bsd boxes doing routing/nat 2014-09-19 10:26:14 <[xming]> okay cleaned up 2014-09-19 10:26:39 its a cable modem, there is nothing i an do about it. 2014-09-19 10:26:58 they dont allow me to put it in bridge mode. 2014-09-19 10:27:05 <[xming]> noway to put it in bridge mode? 2014-09-19 10:27:25 <[xming]> providers who doesn't provide that won't get my money 2014-09-19 10:27:48 its a business contract, which doesnt support it.... duh! 2014-09-19 10:27:48 <[xming]> send try http://sprunge.us/iKWg 2014-09-19 10:28:25 [xming]: i already pushed it 2014-09-19 10:28:32 <[xming]> either they provide me a true modem, or give me a way to bridge 2014-09-19 10:28:40 and added openssl-dev 2014-09-19 10:28:44 <[xming]> clandmeter: ah thanks 2014-09-19 10:29:17 <[xming]> I missed what algitbot said 2014-09-19 10:30:16 I can only choose cable and adsl 2014-09-19 10:30:42 and for adsl im too far away from the station to have a proper connection (even in this dense area) 2014-09-19 10:32:08 im waiting for fiber, if it ever happends... 2014-09-19 10:32:21 <[xming]> (almost) all the providers here in this country have true modems (cable and xDSL) 2014-09-19 10:32:37 <[xming]> maybe I just find that normal 2014-09-19 10:37:54 <[xming]> I can't find the httrack apk, build log looks good, which mirror will be the first to get the uploaded apk? 2014-09-19 10:39:05 <[xming]> http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64/testing/httrack/ 2014-09-19 10:39:14 <[xming]> nothing here http://nl.alpinelinux.org/alpine/v3.0/testing/x86_64/ 2014-09-19 10:46:24 <[xming]> is that real h/w ARM or qemu? 2014-09-19 10:48:28 yes, real hw 2014-09-19 10:48:45 your package has been pushed to edge, not to a stable release 2014-09-19 10:53:22 <[xming]> ah thanks, missed that :/ 2014-09-19 11:30:35 huh 2014-09-19 13:11:30 do somebody know in which package libpci.pc is ? (no in pciutils-dev) 2014-09-19 13:17:55 it should, but i dont see it 2014-09-19 13:18:08 it seems like there is no in dev package 2014-09-19 13:21:43 it's strange, when manually compiling, there is libpci.pc 2014-09-19 13:21:57 <[xming]> not getting packaged? 2014-09-19 13:22:39 nope :\ 2014-09-19 13:26:05 need manual rule to package function for libpci.pc 2014-09-19 13:29:59 http://pastebin.com/ZNZb9nFk if somebody need 2014-09-19 13:32:12 putting it inside for loop isn't the best idea 2014-09-19 13:32:26 your command will be run 4 times 2014-09-19 13:36:06 k0r10n: thanks! 2014-09-19 13:38:42 hi 2014-09-19 13:38:59 are there news because of the wtmp errors 2014-09-19 13:39:01 ? 2014-09-19 13:43:01 clandmeter today i changed my upc contract and i will get new modem, i hope i can setup bridge mode.. There are some: Technicolor TC 7200 ; Thomson TWG870 and mybe Ubee EVW3226 2014-09-19 13:49:57 crow: my modem supports bridge, but they dont allow business customers to use it. 2014-09-19 13:50:33 instead they offer a routed solution which cost 20 euro per month 2014-09-19 13:50:52 but you get 5 usable ips 2014-09-19 14:06:13 i see 2014-09-22 06:24:05 algitbot, welcome back 2014-09-22 07:10:21 morning 2014-09-22 07:13:17 ncopa: any opinion regarding uselessd? 2014-09-22 07:17:23 clandmeter it looks nice, i read about it yesterday, its always good to have alternative. 2014-09-22 07:23:02 clandmeter: not yet 2014-09-22 07:23:10 first impressions though 2014-09-22 07:23:18 1) they are visible in media early 2014-09-22 07:23:33 (might be indication of vaporware) 2014-09-22 07:23:48 2) its a systemd fork 2014-09-22 07:24:01 (so license is same as systemd) 2014-09-22 07:24:09 3) they depend on dbus 2014-09-22 07:24:18 - but no support for kdbus 2014-09-22 07:24:26 afaics 2014-09-22 07:24:44 yeah, that's true 2014-09-22 07:24:53 4) the name is kinda bad :) 2014-09-22 07:24:57 other than that 2014-09-22 07:25:05 looks good 2014-09-22 07:25:30 i like that they want support any /dev manager, mdev, eudev etc 2014-09-22 07:25:41 i like they want be portable 2014-09-22 07:25:51 support for musl 2014-09-22 07:26:07 i like that it makes it possible to reuse systemd unit files 2014-09-22 07:26:14 so yes, this is interesting 2014-09-22 07:26:20 we want keep an eye on it 2014-09-22 07:27:33 yes, the idea is ok, the name is kinda making a statement (which is probably not needed) 2014-09-22 07:29:34 ncopa: i see youve been busy with www. any idea when you want to launch it? 2014-09-22 07:29:45 november 2014-09-22 07:29:48 with v3.1 2014-09-22 07:33:21 would be nice to have a general alpine theme 2014-09-22 07:33:34 then port the theme to the other sites too 2014-09-22 07:33:52 i dunno what to do with forums 2014-09-22 07:33:57 i dont use them myself 2014-09-22 07:35:53 lol 2014-09-22 07:36:03 they sell alpine cdroms 2014-09-22 07:36:05 http://www.zyxware.com/requestcd?distribution_name=Alpine+Linux 2014-09-22 07:37:51 other that i think would be nice is https://bugs.alpinelinux.org/issues/335 2014-09-22 08:37:41 clandmeter: hello, about lxpanel - name of patch missed, should be 0.6.2 not 0.7.0 in APKBUILD 2014-09-22 08:38:12 clandmeter: i thought to rename patch also to 0.7.0 , but did not, sorry 2014-09-22 08:38:52 you will submit a new one or want me to fix it? 2014-09-22 08:39:02 what will be better ? 2014-09-22 08:39:12 the result will be the same. 2014-09-22 08:39:26 Moinmoin 2014-09-22 08:42:05 Is there anything I can do to find the issue on nfs client? 2014-09-22 08:42:56 How can I help you as a not-C-programmer? 2014-09-22 08:43:40 ACTION is @home in the java world...  2014-09-22 08:44:27 clandmeter: http://pastebin.com/S4CGJmLr 2014-09-22 08:44:31 clandmeter: working 2014-09-22 08:59:55 x86 / edge builder is offline for few hours. i'm tracing testing/firefox issues. 2014-09-22 09:05:20 StarWarsFan: i think the problem is rpcbind 2014-09-22 09:05:30 you could try create a backtrace 2014-09-22 09:05:42 hm 2014-09-22 09:05:51 i think i need enable -dbg package for that though 2014-09-22 09:07:31 we have libtirpc-dbg 2014-09-22 09:07:42 k0r10n: if possible can you: git format-patch -1 --stdout | sprunge 2014-09-22 09:17:38 StarWarsFan: apk add libtirpc-dbd rpcbind-dbg; echo 'rc_ulimit="-c unlimited"' >> /etc/conf.d/rpcbind 2014-09-22 09:17:49 then restart rpcbind 2014-09-22 09:18:04 make it segfault and you should have a 'core' file some place 2014-09-22 09:18:22 you can hardcode the place in /proc/sys/kernel/core_pattern 2014-09-22 09:18:42 finally you do: gdb --core /path/to/core /usr/sbin/rpcbind 2014-09-22 09:18:48 and in gdb: bt 2014-09-22 09:18:55 should give you a backtrace 2014-09-22 09:19:16 we have symbols now in musl? 2014-09-22 09:19:26 clandmeter: with apk add musl-dbg yes 2014-09-22 09:19:46 ok nice 2014-09-22 09:19:47 its new? 2014-09-22 09:20:00 maybe 6 months? 2014-09-22 09:20:03 or more 2014-09-22 09:20:16 the -dbg subpkg is probably from last year 2014-09-22 09:20:20 its very nice 2014-09-22 09:20:27 fabled told me we couldnt debug with musl wihtout something added to musl. 2014-09-22 09:20:34 oh 2014-09-22 09:20:38 that i dont know 2014-09-22 09:21:18 he got valgrind working, atleast partially 2014-09-22 09:21:22 last weekend 2014-09-22 09:21:27 which is awesome 2014-09-22 09:21:29 yes i saw something 2014-09-22 09:21:49 i think he wanted to fix his mailbox :) 2014-09-22 09:21:57 yes :) 2014-09-22 09:41:44 re 2014-09-22 09:41:53 ncopa: ok thx, will do this 2014-09-22 09:46:45 ncopa: alpine-iso reports many No such file or directory. Skipped. including stuff for uclibc. do i need to do something different for musl iso? 2014-09-22 09:49:14 clandmeter: maybe git pull? 2014-09-22 09:49:25 i can build isos in musl 2014-09-22 09:49:44 its a fresh pull 2014-09-22 09:49:52 i think the iso is build 2014-09-22 09:50:04 not sure if i can ignore those missing files errors 2014-09-22 09:50:30 is it mkinitfs? 2014-09-22 09:50:35 yes 2014-09-22 09:50:40 http://sprunge.us/iPRQ 2014-09-22 09:51:03 find /etc/mkinitfs -name '*.apk-new' 2014-09-22 09:51:51 make a backup of what you have in /etc/mkinitfs and: for i in $(find /etc/mkinitfs -name '*.apk-new'); do mv $i ${i%.apk-new}; done 2014-09-22 09:52:47 ok, i think i never edited antying in that dir. 2014-09-22 09:58:44 ncopa: thanks, that looks better. 2014-09-22 10:01:42 <3 alpine! 2014-09-22 10:02:02 :) 2014-09-22 10:32:24 <[xming]> apks which are in the edge's testing will be 3.1's main? 2014-09-22 10:33:13 <[xming]> or do I have to do anything to make sure the abuilds I submitted to get into 3.1? 2014-09-22 10:41:16 testing will never go to a release 2014-09-22 10:41:25 you will need to verify its working, and move it to main. 2014-09-22 10:42:34 not sure what the policy is for 3.1, and if all in main will be included. 2014-09-22 11:04:02 [xming], edge/main will get tagged as next release. to get package in testing to main, you need test the package, and request for moving to main with test report (on mailing list or via ticket) 2014-09-22 11:05:32 What does "test the package" mean? Who doesn't test his work before submitting it? :-D 2014-09-22 11:05:38 *his/her 2014-09-22 11:06:44 jomat: sometimes people asks for a package, a developer create it, but does not test 2014-09-22 11:06:59 jomat, we compile test before build 2014-09-22 11:07:07 one is expected to do full usability test before move to main 2014-09-22 11:07:26 to ensure no package mistake, or run-time compatibility issues due to grsec and/or musl 2014-09-22 11:07:29 if we'd only push packages that developer actually use, then we'd not be able to have half the packages 2014-09-22 11:08:49 also, there might be things in the developers environment that is not available on the build server 2014-09-22 11:09:09 so the testing repo is a 'staging' repo to verify that the package built by build server is ok 2014-09-22 13:11:06 nsz: is ncurses-base -> ncurses-terminfo-base ok? in case we get other terminfo sources in future 2014-09-22 13:21:52 i'm probably not the best to give package naming advice :) 2014-09-22 13:22:11 but having terminfo in the name would at least help finding it with apk search 2014-09-22 13:22:41 yeah 2014-09-22 13:22:44 makes sense 2014-09-22 13:51:11 I'm currently creating an APKGBUIL where I want to create a directory owned by the systemuser the package creates. "install -dm755 -o owner ${subpkgdir}/path" doesn't work of course as the owner doesn't exist when the package is packaged. Adding this user to pkgusers doen't work neither, as these users are only created during build-time. Is it the right way to "chown owner /absolute/path" in a pre-install-script? But what happens when the packages is install 2014-09-22 14:20:02 ncopa, do you want us to send you patchs to the new website thru al-dev@al.o or what? 2014-09-22 14:21:43 yes, that is good 2014-09-22 14:22:48 I was thinking that "wiki" "git" "bugs" "forum" could open in a new tab/window 2014-09-22 14:23:20 So the pre-install-script will run chrooted or is there a variable like $rootdir available in them? 2014-09-22 14:24:26 alacerda: i'm not sure i like the 'open-in-new-window' style of links 2014-09-22 14:24:46 but feel free to post a suggestion on ml, might be others prefers it 2014-09-22 14:25:00 jomat: i htink it runs chrooted 2014-09-22 14:25:29 ncopa, ok... I'll send the patch and you comment on it saying that you dont like the idea :) what you think? 2014-09-22 14:25:37 yes 2014-09-22 14:25:40 sounds good 2014-09-22 14:26:11 just sent 2014-09-22 14:26:14 i just pushed script that fetches RSS atom feed 2014-09-22 14:26:36 it shoudl be relatively easy to get feed from bugs.a.o etc 2014-09-22 14:26:38 or blogs 2014-09-22 14:29:12 im not a fan of open-in-new-window links 2014-09-22 14:29:48 just let the user decide on whether or not they want to open the link in a new tab/window 2014-09-22 14:32:25 My patch is loosing the election :^) 2014-09-22 14:32:32 pnutzh4x0r, it makes sense 2014-09-22 14:33:46 yes, i think so too, better let user rightclick and 'open-in-new-tab' 2014-09-22 18:00:20 btw... should a pre-deinstall script try to stop the service that is going to be removed? 2014-09-22 22:12:23 N8@all 2014-09-22 23:38:24 need help building apk. followed the alpine-linux guid but "abuild package" does not create any *.apk file 2014-09-22 23:42:49 package Create package in \$PKGDEST -> but no pkg is created ... did i miss something? 2014-09-22 23:46:14 nobody? 2014-09-23 05:57:30 morning 2014-09-23 05:58:19 jomat: we have normally not started/stopped anything from install/deinstall scripts 2014-09-23 06:01:44 we could maybe consider check if service is running on pre-deinstall and exit with error if it is 2014-09-23 07:45:16 ncopa: do we have telnetd somewhere? 2014-09-23 07:46:12 seems bb is build without it 2014-09-23 07:47:51 hm 2014-09-23 07:47:54 i dont think so 2014-09-23 07:48:31 CONFIG_TELNET=y 2014-09-23 07:48:31 CONFIG_FEATURE_TELNET_TTYPE=y 2014-09-23 07:48:32 CONFIG_FEATURE_TELNET_AUTOLOGIN=y 2014-09-23 07:48:32 # CONFIG_FEATURE_TELNETD_STANDALONE is not set 2014-09-23 07:48:32 # CONFIG_TELNETD is not set 2014-09-23 07:48:32 # CONFIG_FEATURE_TELNETD_INETD_WAIT is not set 2014-09-23 07:55:52 btw, I have enabled the git feed on wwwtest.alpinelinux.org 2014-09-23 07:56:11 note how fast the "Latest Development" is updated after you do gi push ;) 2014-09-23 07:56:14 git* 2014-09-23 08:07:18 there is a drop shadow under the footer: http://wwwtest.alpinelinux.org/downloads/ 2014-09-23 08:07:43 but it is not visible if the page fills up the browser window 2014-09-23 08:08:21 try resize the window til scrollbar appears 2014-09-23 08:08:30 then will the drop shadow not be visible 2014-09-23 08:08:48 should i add a small margin under so that the dropshadow becomes visible when scrolling too? 2014-09-23 08:16:16 <[xming]> morning, some publicity for alpine :D 2014-09-23 08:16:18 <[xming]> http://linux.slashdot.org/story/14/09/22/2245217/outlining-thin-linux?utm_source=rss1.0mainlinkanon&utm_medium=feed 2014-09-23 08:18:19 cool :) 2014-09-23 08:25:01 <[xming]> patch for testing/irrlicht http://sprunge.us/NKgE 2014-09-23 08:25:32 <[xming]> mv headers from /usr/include/irrlicht/include to /usr/include/irrlicht 2014-09-23 08:27:30 it seems incorrect 2014-09-23 08:27:37 ah, ok, I cannot read 2014-09-23 08:27:43 <[xming]> it's a mess 2014-09-23 08:27:57 <[xming]> but I didn't want to change the APKBUILD much 2014-09-23 08:28:58 <[xming]> and .so is still versioned @ 1.8.0 :/ 2014-09-23 08:29:21 irrlicht devs doesn't really seem to care about non-windows users 2014-09-23 08:29:22 <[xming]> but it's the same on my Gentoo box, so I presume it's upstream 2014-09-23 08:29:56 let my hard drive unchoke after entering aports tree and I'll push it 2014-09-23 08:30:17 <[xming]> I have apkbuild for minetest (which contain the client and common) and minetest-server 2014-09-23 08:30:27 <[xming]> barthalion: thanks 2014-09-23 08:31:07 <[xming]> need to fix them after irrlicht's include path is fixed 2014-09-23 08:32:02 <[xming]> what kind of h/w is the ARM buildbot running on? 2014-09-23 08:36:09 <[xming]> does alpine has something like this? http://euscan.iksaif.net/ 2014-09-23 08:37:26 <[xming]> barthalion: minetest's wiki claims that levedb is more reliable than sqlite, so including sqlite for client might be worthwile 2014-09-23 08:38:01 I really doubt in any noticable difference, but it's up to you 2014-09-23 08:38:29 about euscan… ncopa uses a script to compare Alpine repositories with Arch, probably he added some other repos too 2014-09-23 08:38:54 I wrote similiar script too, but I'm not packaging for Alpine lately 2014-09-23 08:39:26 <[xming]> barthalion: up to me? :D But I can't decide :D 2014-09-23 08:40:20 <[xming]> I will leave the leveldb out, and let the users decide, I am only interesting in the server 2014-09-23 08:41:31 ncopa: I asked because xdm left its /var/run/xdm.pid on my build host when I first removed the package and then killed the daemon 2014-09-23 08:42:58 jomat: i understand the problem with not properly cleaning up 2014-09-23 08:43:21 i would prefer not start/stop anything from apk scripts though 2014-09-23 08:43:39 we could have the pre-script prevent the package form be deleted though 2014-09-23 08:43:48 so you have to manually stop it before you can apk del it 2014-09-23 08:43:59 That sounds good! 2014-09-23 08:44:31 Packages starting their daemons are nightmares (didn't we have that yesterday, too? :-D ) 2014-09-23 08:46:26 <[xming]> hmm minetest and killing daemons hmmm 2014-09-23 08:48:02 i wonder if we could have some apk support for it 2014-09-23 08:48:16 if there is a file in /etc/init.d/ 2014-09-23 08:48:44 then if rc-status returns that it is running, then exit with error 2014-09-23 08:49:24 then we dont need write any pre-deinstall scripts at all 2014-09-23 08:49:41 <[xming]> but killing daemons is *fun* 2014-09-23 08:49:53 similar to trigger, but reverse, for deinstall 2014-09-23 08:50:40 <[xming]> that's nice, gentoo misses that 2014-09-23 08:51:03 fabled: any thoughts about that? 2014-09-23 08:51:23 the idea is that openrc ships a "deltrigger" that monitors /etc/init.d 2014-09-23 08:51:43 the trigger is executed before anything is actually deleted 2014-09-23 08:51:55 and the triger must return success, or the apk del operation is aborted 2014-09-23 08:52:08 we have similar ticket 2014-09-23 08:52:11 on restarting services 2014-09-23 08:52:35 I occasionally get things like wget: bad address 'nl.alpinelinux.org' 2014-09-23 08:53:02 basically making me unable to do anything other than apk update 2014-09-23 08:54:03 <[xming]> I think I had that once yesterday 2014-09-23 08:54:35 and now I can't get rid of it during docker build 2014-09-23 08:54:38 dns issue? 2014-09-23 08:55:09 probably, however apk update doesn't fail 2014-09-23 08:55:11 just wget 2014-09-23 08:55:26 ncopa, http://bugs.alpinelinux.org/issues/2251 2014-09-23 08:55:35 barthalion: what happens if you do wget manually? 2014-09-23 08:55:58 basically restarting service on package upgrade - and refusal to delete if service is running - share the same logic/trigger framework 2014-09-23 08:56:10 let me check 2014-09-23 08:56:29 ugh, it worked 2014-09-23 08:56:34 or not 2014-09-23 08:56:37 yeah, same thing 2014-09-23 08:57:52 uggedal is MIA? 2014-09-23 08:58:21 fabled: hmmm 2014-09-23 08:59:09 barthalion: seems like he moved over to void linux 2014-09-23 08:59:25 The .org nameservers know a ns3.alpinelinux.org (213.234.126.134) which does not respond 2014-09-23 08:59:41 ah... jomat, I'll fix that 2014-09-23 09:00:00 maybe file a bug on bugs.a.o under 'infra' 2014-09-23 09:01:14 :/ 2014-09-23 09:01:18 I'll purge docker cache and see what happens then 2014-09-23 09:09:45 seems to be working now 2014-09-23 09:09:47 docker magic 2014-09-23 09:15:31 <[xming]> I looked into docker a few weeks back, it's nice for app isolation, but a bit too bloated and not made for full container management IMHO 2014-09-23 09:21:50 <[xming]> grrr irrlicht-dev is still wrong 2014-09-23 09:22:23 <[xming]> when I did "abuild rootpkg" and checked the pkg/irrlicht-dev it all seemed correct 2014-09-23 09:22:44 <[xming]> but now the header files are still in the wrong place 2014-09-23 09:28:34 what is the correct place? 2014-09-23 09:34:49 ugh 2014-09-23 09:35:03 what is /usr/share/$pkgname/examples/bin ? 2014-09-23 09:35:19 that looks wrong... 2014-09-23 09:35:39 the /usr/share is for arch independent data 2014-09-23 09:37:37 <[xming]> idd, it's a messy apkbuild, maybe I should just start from scratch 2014-09-23 09:37:55 <[xming]> headers get into /usr/include/irrlicht/include 2014-09-23 09:38:12 <[xming]> should be /usr/include/irrlicht 2014-09-23 09:39:36 <[xming]> how so I manipulate files in the subpackage? 2014-09-23 09:40:24 <[xming]> tried dev() but obviously that overwrite the default func and I don't get any files there 2014-09-23 09:40:32 you can start from scratch if you want 2014-09-23 09:41:04 <[xming]> irrlicht it's a mess :( not just he apkbuild 2014-09-23 09:41:28 <[xming]> I don't think that those demo bins benifits anyone 2014-09-23 09:41:44 <[xming]> if anyone want them they can compile those 2014-09-23 09:41:50 if you want to reuse that APKBUIL, I'd start with something like: https://gist.github.com/ncopa/3603bb5c68451e61da2f 2014-09-23 09:42:00 and then try get rid of all the cd ../.. 2014-09-23 09:42:05 <[xming]> or should those bins in a subpkg like examples? 2014-09-23 09:42:29 yes, if we want ship them, they yes, lets put them in a subpkg 2014-09-23 09:42:54 but I think we can just drop them 2014-09-23 09:43:23 http://pkgs.fedoraproject.org/cgit/irrlicht.git/tree/irrlicht.spec 2014-09-23 09:43:45 they dont seem to ship the examples 2014-09-23 09:44:16 <[xming]> okay I'll skip them 2014-09-23 09:44:23 but they do ship a subpackage for irrXML 2014-09-23 09:45:44 i'm ok with calling it irrlicht-xml 2014-09-23 09:46:31 <[xming]> gentoo doesn't install those example bins either 2014-09-23 09:47:16 <[xming]> do you want me to make an extra subpkg for the xml? 2014-09-23 09:47:31 <[xming]> or just irrlicht that bundels all of them? 2014-09-23 09:48:10 <[xming]> I don't agree that fedora ships libIrrXML.so with irrXML-dev 2014-09-23 09:48:23 <[xming]> dynamic libs do not belong in -dev 2014-09-23 09:49:36 its a symlink right? 2014-09-23 09:49:41 the symlink belong in -dev 2014-09-23 09:49:57 and we normally dont split to various -dev packages 2014-09-23 09:50:08 so we dont do irrXML-dev 2014-09-23 09:50:23 if you want use it, then add irrlicht-dev to makedpends 2014-09-23 09:51:01 basically, the -dev packages pulls in all the subpackages 2014-09-23 09:51:16 so it corresponds more or less to the corresponding gentoo package 2014-09-23 09:51:25 (which default installs headers etc) 2014-09-23 09:52:09 <[xming]> I don't see irrlicht-xml in gentoo, and I dont see it built by default 2014-09-23 09:52:24 heh 2014-09-23 09:52:29 then dont build it 2014-09-23 09:52:30 <[xming]> ln -s libIrrXML.so.%{irrlicht_version} libIrrXML.so.1 2014-09-23 09:52:36 <[xming]> that's fedora 2014-09-23 09:52:49 if nothing uses/needs irrXML, then dont build it 2014-09-23 09:52:57 <[xming]> okay 2014-09-23 10:06:11 ACTION hides 2014-09-23 10:41:50 <[xming]> manually cp/install is way better than the make install of irrlicht 2014-09-23 11:24:21 <[xming]> why does libIrrlicht.so get in irrlicht-dev package and not like libIrrlicht.so.1.8 which is in irrlicht (http://sprunge.us/GLKG) 2014-09-23 11:24:45 [xming]: it is a symlink right? 2014-09-23 11:25:08 <[xming]> ncopa: yes 2014-09-23 11:25:11 readelf -d shows the SONAME 2014-09-23 11:25:28 should be something like libIrrlicht.so.1 or similar 2014-09-23 11:25:47 that is what the libc's dynamic linker will look for in applications that linked to it 2014-09-23 11:26:02 the .so symlink is only needed build time 2014-09-23 11:26:38 gcc -lIrrlicht will look for libIrrlicht.so 2014-09-23 11:26:51 <[xming]> (SONAME) Library soname: [libIrrlicht.so.1.8] 2014-09-23 11:26:57 correct 2014-09-23 11:27:01 so during build 2014-09-23 11:27:07 <[xming]> which is a symlink that I made 2014-09-23 11:27:32 you need the .so symlink to be able to do gcc -lIrrlicht 2014-09-23 11:27:44 <[xming]> ah 2014-09-23 11:27:57 but you dont need that .so runtime 2014-09-23 11:28:09 what you need is specified in NEEDED 2014-09-23 11:28:15 so 2014-09-23 11:28:17 <[xming]> now I get it and somehow vaguely remembers that 2014-09-23 11:28:24 gcc -lIrrlicht -o myapp 2014-09-23 11:28:43 readelf -d myapp will have NEED libIrrlicht.so.1.8 2014-09-23 11:29:06 <[xming]> yes :D 2014-09-23 11:29:07 where libIrrlicht.so.1.8 is the SONAME found during gcc -lIrrlicht 2014-09-23 11:29:33 <[xming]> but gcc needs .so for -lIrrlicht os it's in the -dev 2014-09-23 11:29:40 correct 2014-09-23 11:29:58 other binary distros does similar 2014-09-23 11:30:01 <[xming]> totally forgot all this kind of stuff after 10 years working on gentoo 2014-09-23 11:30:08 yeah 2014-09-23 11:30:20 gentoo does not split -dev and runtime packages 2014-09-23 11:35:40 <[xming]> Maintainer: Jeff Bilyk 2014-09-23 11:35:44 <[xming]> is he still active? 2014-09-23 11:36:07 he does infra afaik 2014-09-23 11:36:21 i dont think he minds if you take over the package 2014-09-23 11:38:03 <[xming]> or I just add myself as contributor? 2014-09-23 11:38:46 yeah 2014-09-23 11:48:17 <[xming]> okay finally r2 of irrlicht 2014-09-23 11:48:20 <[xming]> http://sprunge.us/RGeg 2014-09-23 11:49:14 <[xming]> minetest build nicely against r2 2014-09-23 11:49:58 <[xming]> so AFAI am concerned it's good :p 2014-09-23 12:05:29 [xming]: looks good to me 2014-09-23 12:05:35 i wonder if we can drop the .a file 2014-09-23 12:05:46 + install -m644 "$_builddir"/lib/Linux/libIrrlicht.a "$pkgdir"/usr/lib 2014-09-23 12:05:55 and only provide dynamic lib 2014-09-23 12:06:05 <[xming]> I think we can, notone wants to link statically to it IMHO 2014-09-23 12:06:09 <[xming]> it's so huge 2014-09-23 12:06:35 http://sprunge.us/KKDT 2014-09-23 12:06:49 like that? 2014-09-23 12:06:57 i just commit --amend that 2014-09-23 12:07:02 <[xming]> yeah 2014-09-23 12:08:14 you dont need add musl-dev to makedepends 2014-09-23 12:08:26 its part of build-base 2014-09-23 12:08:39 just like you dont need add gcc or make there 2014-09-23 12:09:01 <[xming]> okay :D 2014-09-23 12:09:06 i fix that too 2014-09-23 12:09:14 <[xming]> wait a moment 2014-09-23 12:09:19 <[xming]> before you commit 2014-09-23 12:09:49 <[xming]> if you drop the .a you can also drop the building of it 2014-09-23 12:09:52 <[xming]> make || return 1 2014-09-23 12:09:58 ok 2014-09-23 12:10:05 <[xming]> only this is needed 2014-09-23 12:10:07 <[xming]> make sharedlib || return 1 2014-09-23 12:10:14 good 2014-09-23 12:10:26 i wonder if the headers in -dev package does #include 2014-09-23 12:10:43 where is provided in some of the makedepends 2014-09-23 12:11:52 <[xming]> I have build minetest against r2, and it's usccessful 2014-09-23 12:12:07 <[xming]> will push minetest and minetest-server later (or tomorrow) 2014-09-23 12:12:12 #include 2014-09-23 12:12:34 nah 2014-09-23 12:12:40 xbox/windows stuff 2014-09-23 12:12:41 ok 2014-09-23 12:12:43 looks good 2014-09-23 12:14:12 here we go 2014-09-23 12:14:13 thanks! 2014-09-23 12:16:19 <[xming]> I meant it was successful 2014-09-23 12:16:41 <[xming]> usuccesful is ambiguous :p 2014-09-23 13:21:01 should probably update rpi tarball at somepoint 2014-09-23 13:21:07 got the new kernel out now 2014-09-23 18:55:02 hi@all 2014-09-23 18:55:49 i'm trying to reproduce the nfs-problem right now 2014-09-23 18:56:15 so i setup two virtual machines based on alpine-mini-3.0.4-x86_64.iso 2014-09-23 18:56:36 then i tried to setup nfs-server like on the wiki page: 2014-09-23 18:56:41 http://wiki.alpinelinux.org/wiki/Setting_up_a_nfs-server 2014-09-23 18:57:35 but: 2014-09-23 18:57:37 nfs-server:~# rc-service nfs start 2014-09-23 18:57:37 * Exporting NFS directories ... [ ok ] 2014-09-23 18:57:37 * Starting NFS mountd ... [ ok ] 2014-09-23 18:57:37 * Starting NFS daemon ... 2014-09-23 18:57:37 rpc.nfsd: writing fd to kernel failed: errno 111 (Connection refused) 2014-09-23 18:57:39 rpc.nfsd: unable to set any sockets for nfsd [ !! ] 2014-09-23 18:57:41 * Starting NFS smnotify ... [ ok ] 2014-09-23 18:57:44 * ERROR: nfs failed to start 2014-09-23 18:57:46 nfs-server:~# 2014-09-23 19:44:03 StarWarsFan: what kind of virtual machine? 2014-09-23 19:44:23 VMware workstation 9 2014-09-23 19:44:53 ok, that should be ok 2014-09-23 19:45:55 trying the same with alpine 2.7 right now... 2014-09-23 19:47:44 2.7 works fine for me 2014-09-23 19:48:48 StarWarsFan: which kernel? 2014-09-23 19:49:04 hm, same on 2.7 :-( 2014-09-23 19:49:18 nfs-server-2-7:~# uname -a 2014-09-23 19:49:18 Linux nfs-server-2-7 3.10.52-0-grsec #1-Alpine SMP Tue Aug 12 06:08:58 UTC 2014 x86_64 Linux 2014-09-23 19:50:02 nfs-server-2-7:~# rc-service nfs start 2014-09-23 19:50:02 * Mounting nfsd filesystem in /proc ... [ ok ] 2014-09-23 19:50:02 * Exporting NFS directories ... [ ok ] 2014-09-23 19:50:02 * Starting NFS mountd ... [ !! ] 2014-09-23 19:50:02 * Starting NFS daemon ... 2014-09-23 19:50:03 rpc.nfsd: writing fd to kernel failed: errno 111 (Connection refused) 2014-09-23 19:50:05 rpc.nfsd: unable to set any sockets for nfsd [ !! ] 2014-09-23 19:50:07 * Starting NFS smnotify ... [ ok ] 2014-09-23 19:50:10 * ERROR: nfs failed to start 2014-09-23 19:51:24 this is what i get 2014-09-23 19:51:25 http://sprunge.us/GYgS 2014-09-23 19:51:49 and in dmesg: [1598845.851376] rpcbind[6768]: segfault at 2abf57b0 ip 00007f3f2abe65cd sp 00007fff07392708 error 4 in ld-musl-x86_64.so.1[7f3f2ab97000+85000] 2014-09-23 19:53:03 now i'm completely confused as it is even on 2.7 _not_ working... 2014-09-23 19:54:19 wasnt that the issue you were after? 2014-09-23 19:54:27 the segfault in rpcbind? 2014-09-23 19:58:23 yes, but it is working on my current systems with 2.7 2014-09-23 19:58:28 only on 3.0-64 not 2014-09-23 19:58:44 but now, as i want to reproduce the problem, nfs is not starting even on 2.7 2014-09-23 19:58:54 so what the heck is wrong!? 2014-09-23 19:59:36 well its working on my home box 2014-09-23 20:01:08 do you have an idea? i installed a clean new mini-iso 2014-09-23 20:01:14 then apk add nfs-utils 2014-09-23 20:01:19 and it is not starting 2014-09-23 20:02:44 StarWarsFan: just installed 2.7 2014-09-23 20:02:46 works fine 2014-09-23 20:03:29 what tells uname -a 2014-09-23 20:04:07 http://sprunge.us/RahQ 2014-09-23 20:04:29 Linux localhost 3.10.52-0-grsec #1-Alpine SMP Tue Aug 12 06:08:58 UTC 2014 x86_64 Linux 2014-09-23 20:04:57 unbelievable 2014-09-23 20:05:04 so what is the difference? 2014-09-23 20:05:21 not much 2014-09-23 20:05:25 this is also vmware 2014-09-23 20:06:17 StarWarsFan: can you sprunge your dmesg on 2.7? 2014-09-23 20:06:31 http://sprunge.us/iLSj 2014-09-23 20:09:14 http://sprunge.us/DSGd 2014-09-23 20:09:30 your 2.7 error is diff then your 3.0 error 2014-09-23 20:09:45 i guess you also have the segfault on 3.0 2014-09-23 20:14:46 this is from 3.0 nfs-client: http://sprunge.us/gdXK 2014-09-23 20:18:28 clandmeter: did you tried it on vmware workstation too or somewhere else? 2014-09-23 20:30:38 esxi 2014-09-23 20:32:23 StarWarsFan: the only diff could be networking 2014-09-23 20:33:32 im unable to debug rpc in gdb 2014-09-23 20:33:39 it complains about crc mismatch 2014-09-23 20:35:03 wtf 2014-09-23 20:35:09 ah, got a hint! 2014-09-23 20:35:19 i rebooted, and now nfs works on 3.0 2014-09-23 20:35:29 i reinstalled the whole system using alpine-2.7.9-x86_64.iso and nfs is starting 2014-09-23 20:35:39 if i use the mini-iso, it fails 2014-09-23 20:36:46 that makes no sence, it will install the system from aports 2014-09-23 20:36:54 right 2014-09-23 20:37:04 but this is what i get right now :-/ 2014-09-23 20:37:13 will verify again be reinstalling the mini-iso... 2014-09-23 20:38:59 im unable to segfault it now. 2014-09-23 20:39:03 dunno why not. 2014-09-23 20:41:50 ok, forget it, after reinstalling the mini-iso again, nfs is starting on 2.7 now 2014-09-23 20:42:02 wtf is going on here!? 2014-09-23 20:42:07 confusing, very confusing... 2014-09-23 20:42:22 will go ahead on 3.0 now... 2014-09-23 21:19:10 i give up for today... 2014-09-23 21:19:13 n8@all 2014-09-24 05:41:49 morning 2014-09-24 07:04:07 fabled: i was trying gdb yesterday, and it was complaining about crc mismatch between regular pkg and dbg pkg. did i do something wrong? 2014-09-24 07:21:24 clandmeter, are they identical versions? perhaps pkg was updated, and you have old main pkg, and new -dbg? 2014-09-24 07:21:47 thats what i thought, but it was matching. 2014-09-24 07:22:41 I didnt test any furter, so it could be i didnt someting wrong. If i have it again ill let you know. 2014-09-24 08:24:54 <[xming]> ouch we don't have leveldb? 2014-09-24 08:44:54 hello all 2014-09-24 08:45:38 you got a friend 2014-09-24 08:45:58 hi :) 2014-09-24 08:46:14 :) 2014-09-24 08:46:26 do somebody know about TEMP_FAILURE_RETRY ? 2014-09-24 08:46:41 i've got som errors while compiling 2014-09-24 08:46:53 it's in unistd.h in glibc linux distro 2014-09-24 08:47:10 does musl have a replace for that? 2014-09-24 08:47:44 i dunno, maybe better to ask in #musl. 2014-09-24 08:48:09 ok, will try, tyvm 2014-09-24 08:48:22 although maybe fabled here knows about it, but he is also active in #musl 2014-09-24 08:48:39 hmm, ok 2014-09-24 10:45:32 when moving to musl from uclibc what main libc lib gets changed in /lib dir? 2014-09-24 10:45:46 sorry not yet installed v3.0 2014-09-24 10:57:35 are any apps still being compiled with uclibc(statically) in v3.0? 2014-09-24 10:59:31 the musl dynamic loader is /lib/ld-musl-x86_64.so.1 and libc is referenced in shared libraries as libc.musl-x86_64.so.1 which is a symlink in /lib to the loader 2014-09-24 10:59:56 s/x86_64/i386/g on i386 2014-09-24 11:00:31 thanks 2014-09-24 11:01:47 if the loader is set to the musl one in the elf header of an executable then musl will be used (the loader has the whole libc in it) 2014-09-24 11:03:13 (so all libc.* libm.* etc shared object dependencies will be skipped at load time and implicitly work) 2014-09-24 11:06:05 just noticed commits e619a482fcbd7421 , 01907687add4bd2c and was wondering if any other libs got deleted while moving to musl, and not restored similarly 2014-09-24 11:08:03 commit 38bbb840dad79 2014-09-24 14:19:01 files from v2.5.4-342-g9b36734 uploaded 2014-09-24 14:19:11 webkitgtk2 was added... 2014-09-24 14:19:43 files from v2.5.4-342-g9b36734 uploaded 2014-09-24 14:47:17 \o/ 2014-09-24 14:49:04 pnutzh4x0r: sorry that i forgot that patch... 2014-09-24 14:50:15 ncopa: no problem 2014-09-24 14:50:38 i have one more cleanup coming 2014-09-24 14:54:20 ah, i didnt know checkpath existed 2014-09-24 14:54:23 i can use that from now on 2014-09-24 14:54:30 i had to modify the mpd init as well 2014-09-24 15:07:20 ncopa: i had to patch one of the cmake scripts in weechat just now to get it to compile 2014-09-24 15:08:07 moskvax: can you apk add sprunge && git diff . | sprunge 2014-09-24 15:08:14 oh 2014-09-24 15:08:20 you had to patch cmake scritps 2014-09-24 15:08:24 ncopa: it's due to libintl 2014-09-24 15:08:30 ncopa: and that musl has ngettext 2014-09-24 15:09:00 ncopa: it leads to a linker error because the build script doesn't link libintl in explicitly 2014-09-24 15:09:25 sounds like a not too uncommon error 2014-09-24 15:10:09 ncopa: it includes the gnu libintl header but not the gnu library 2014-09-24 15:10:50 ok 2014-09-24 15:10:54 ncopa: should i make an issue on the tracker for it & upload the patch? 2014-09-24 15:11:15 yes that would be nice 2014-09-24 15:11:21 i need to run now 2014-09-24 15:11:33 ok, i'll do that now 2014-09-24 15:11:35 the tracker helps me not for get it 2014-09-24 15:11:36 thanks! 2014-09-24 15:12:15 np :) 2014-09-24 15:29:46 ah this is really because of a conflict with gettext-dev which makes libintl.h 2014-09-24 15:30:39 libintl.h should probably be in libintl package and not in gettext-dev 2014-09-24 16:31:19 hey all i am taking a look at alpine development 2014-09-24 16:31:50 and I am not seeming to find where to build alpine including the kernel from source 2014-09-24 16:32:12 *am not finding where 2014-09-24 16:47:02 systmkor: http://wiki.alpinelinux.org/wiki/Developer_Documentation 2014-09-24 16:47:10 systmkor: http://git.alpinelinux.org/cgit/aports/ 2014-09-24 16:50:34 pnutzh4x0r: my b, i thought aports was for only building specific ports/packages on alpine 2014-09-24 20:38:22 n8@all 2014-09-25 04:34:08 is there any formalized layout of alpine mirrors? 2014-09-25 04:34:30 such as having requiring a ".latest.x86.txt" file 2014-09-25 06:36:37 mornings 2014-09-25 06:36:39 huh 2014-09-25 06:36:48 morning 2014-09-25 06:36:55 Installed: Available: 2014-09-25 06:36:55 bash-4.3.025-r0 = 4.3.025-r0 2014-09-25 06:37:12 $ env x='() { :;}; echo vulnerable' /bin/bash -c "echo this is a 2014-09-25 06:37:12 test" 2014-09-25 06:37:12 vulnerable 2014-09-25 06:37:12 this is a test 2014-09-25 06:38:17 <[xming]> moin 2014-09-25 06:41:24 looks like 4.3.025 does not fix it? 2014-09-25 06:41:29 <[xming]> time to bash bash :p 2014-09-25 06:41:34 yeah 2014-09-25 06:44:49 <[xming]> should be 2014-09-25 06:44:59 <[xming]> 4.2.024 is fixed 2014-09-25 06:45:09 <[xming]> 4.2.048 is fixed 2014-09-25 06:45:32 <[xming]> http://article.gmane.org/gmane.comp.security.oss.general/13852 2014-09-25 06:46:57 <[xming]> env x='() { :;}; echo vulnerable' bash -c "echo this is a test" 2014-09-25 06:47:04 <[xming]> bash: warning: x: ignoring function definition attempt 2014-09-25 06:47:12 <[xming]> bash: error importing function definition for `x' 2014-09-25 06:47:18 <[xming]> this is a test 2014-09-25 06:48:09 ncopa-desktop:~$ env x='() { :;}; echo vulnerable' bash -c "echo" 2014-09-25 06:48:09 vulnerable 2014-09-25 06:48:53 ncdev-edge-x86_64:~/aports/main/bash$ env x='() { :;}; echo vulnerable' bash -c 2014-09-25 06:48:53 "echo this is a test" 2014-09-25 06:48:53 vulnerable 2014-09-25 06:48:53 this is a test 2014-09-25 06:48:58 <[xming]> $ bash --version 2014-09-25 06:49:05 <[xming]> GNU bash, version 4.2.48(1)-release (x86_64-pc-linux-gnu) 2014-09-25 06:49:13 ncdev-edge-x86_64:~/aports/main/bash$ bash --version 2014-09-25 06:49:13 GNU bash, version 4.3.0(1)-release (x86_64-alpine-linux-musl) 2014-09-25 06:49:13 Copyright (C) 2013 Free Software Foundation, Inc. 2014-09-25 06:49:13 License GPLv3+: GNU GPL version 3 or later 2014-09-25 06:49:13 This is free software; you are free to change and redistribute it. 2014-09-25 06:49:14 There is NO WARRANTY, to the extent permitted by law. 2014-09-25 06:49:20 looks like patches are not applied 2014-09-25 06:49:40 <[xming]> there you go :D 2014-09-25 06:50:51 i think you are right 2014-09-25 06:50:54 didnt look at it 2014-09-25 06:51:01 i think thats my mistake 2014-09-25 06:51:08 I'll fix it *now* 2014-09-25 06:51:14 none of them are applied 2014-09-25 06:51:24 except the regular ones 2014-09-25 06:52:31 like any good pacakger, one should test the result of his product. i guess it was too late for me :) 2014-09-25 06:55:34 $source.patch::url should fix it. or just change the patch funct. 2014-09-25 06:56:28 <[xming]> easy say then done 2014-09-25 06:56:56 <[xming]> we all know that and still not do that every time 2014-09-25 06:58:30 ugh 2014-09-25 06:58:38 actually somebody has played with it. it should have been obvious it was broken, even for me. 2014-09-25 06:59:01 i think i have applied sec fixes earlier to bash 2014-09-25 06:59:04 that never was applied... 2014-09-25 06:59:06 :-( 2014-09-25 06:59:08 ok 2014-09-25 06:59:15 <[xming]> oh joy 2014-09-25 06:59:18 this was actually a bit easy to get wrong.. 2014-09-25 06:59:34 or it was not that obvious how to do it right 2014-09-25 06:59:37 <[xming]> thank god I don't use bash in my alpine containers :p 2014-09-25 06:59:49 i think i need refactor it a bit 2014-09-25 07:00:13 yeah 2014-09-25 07:00:30 well last sec issue was debated if it was a sec issue at all 2014-09-25 07:00:51 was non critical 2014-09-25 07:00:55 this one otoh... 2014-09-25 07:00:58 <[xming]> but I do use a lot of bash else where :( 2014-09-25 07:01:27 <[xming]> I am a #!/bin/bash junkie 2014-09-25 07:01:33 yeah, bash is way too popular 2014-09-25 07:01:49 bash is too big and slow for my taste 2014-09-25 07:01:53 <[xming]> I actually do use a lof of bashism 2014-09-25 07:01:58 <[xming]> idd it's slow 2014-09-25 07:02:05 <[xming]> and bloated 2014-09-25 07:02:15 they even admit it in the man page 2014-09-25 07:02:19 BUGS section 2014-09-25 07:02:22 its too big and slow 2014-09-25 07:02:38 lua is like half the size of bash 2014-09-25 07:03:00 <[xming]> that's one of the reason I want to switch to alpine for VMs and containers 2014-09-25 07:03:16 lua has better array/hashtable implementation 2014-09-25 07:03:17 <[xming]> gentoo defaults to bash 2014-09-25 07:03:20 and is *way* faster 2014-09-25 07:03:26 i know 2014-09-25 07:03:38 <[xming]> I tried lua, but can't get my head around those metatables 2014-09-25 07:03:51 bash was one of the motivators for me to abandon gentoo 2014-09-25 07:04:06 <[xming]> now fell in love with Golang, but it's not scripting 2014-09-25 07:04:49 <[xming]> Gentoo is falling behind, it's not dying yet but it's not growing for sure 2014-09-25 07:05:17 <[xming]> used it since (I think) 2002, after LFS 2014-09-25 07:05:49 <[xming]> Gentoo is not even bleeding edge any more 2014-09-25 07:09:41 <[xming]> to bash or not to bash \o/ 2014-09-25 07:10:26 lol 2014-09-25 07:10:29 GNU bash, version 4.3.0(1)-release-(x86_64-alpine-linux-musl) is that the correct patch level? 2014-09-25 07:10:36 its wrong 2014-09-25 07:10:39 i think i just fixed 2014-09-25 07:10:47 but seems like v2.7 and older are ok 2014-09-25 07:10:53 i apk updated 2014-09-25 07:11:24 ncopa-desktop:~$ bash --version 2014-09-25 07:11:24 GNU bash, version 4.3.25(1)-release (x86_64-alpine-linux-musl) 2014-09-25 07:11:41 maybe it installed a local build? 2014-09-25 07:12:00 apk info -v |grep bash 2014-09-25 07:12:01 bash-4.3.025-r2 2014-09-25 07:12:35 oh i know already 2014-09-25 07:12:44 its my local version 2014-09-25 07:13:09 abuild cleanoldpkg 2014-09-25 07:13:12 should remove it 2014-09-25 07:15:35 it does remove it, but local apkindex still has r2. 2014-09-25 07:15:51 is there a abuild cmd to updte apkindex? 2014-09-25 07:16:20 yes, 2014-09-25 07:16:41 dont tell me its abuild -r :) 2014-09-25 07:16:46 abuild update_abuildrepo_index 2014-09-25 07:16:54 its not that obvious 2014-09-25 07:17:28 maybe it can be exec with cleanoldpkg? 2014-09-25 07:17:29 ncopa, could we just have it as 'abuild index' ? 2014-09-25 07:17:39 maybe we should have a short alias 2014-09-25 07:17:41 yes 2014-09-25 07:17:44 abuild index 2014-09-25 07:17:59 and we should likelky call it with cleanpkg and cleanoldpkg 2014-09-25 07:23:00 files from v2.5.4-343-gad7e76b uploaded 2014-09-25 07:54:14 was there anything else we needed to be fixed in abuild? 2014-09-25 09:08:19 btw, anyone is working on google chromium? 2014-09-25 09:08:37 i tested netflix + google chrome on ubuntu yesterday 2014-09-25 09:08:40 works great 2014-09-25 09:08:46 woudl be cool to get that working on alpine 2014-09-25 09:48:19 is there anyone working on packaging chromium web browser? 2014-09-25 10:20:05 clandmeter, you broke modloop 2014-09-25 10:20:34 + for modloop in ${alpine_mnt}$KOPT_modloop \ 2014-09-25 10:20:35 + ${alpine_mnt}$KOPT_BOOT_IMAGE.modloop.* \ 2014-09-25 10:20:35 + ${alpine_mnt}$KOPT_BOOT_IMAGE.cmg; do 2014-09-25 10:20:35 + [ -f "$modloop" ] || continue 2014-09-25 10:20:35 + done 2014-09-25 10:20:42 it never breaks on found image 2014-09-25 10:21:45 so the correct modloop is not found, and starting it fails 2014-09-25 10:40:43 sounds like -- [ -f "$modloop" ] && break -- was intended 2014-09-25 10:45:30 fabled: yes, i think i should replaced it with break. 2014-09-25 11:06:44 ncopa: i'm working on chromium 2014-09-25 11:23:18 k0r10n: so v8 is for chromium? 2014-09-25 11:30:01 ncopa: yes :) 2014-09-25 11:32:29 ncopa: i also got ninja and gtest for building, but thinking about using simple make (but ninja is very fast) 2014-09-25 11:49:05 seems nice http://www.sysdig.org/ 2014-09-25 11:49:31 monit v5.9, update pls 2014-09-25 12:00:33 k0r10n: i got significantly more interested in chromium after this netflix support... :) 2014-09-25 12:02:59 ncopa: :):) 2014-09-25 12:06:14 clandmeter, care to fix the modloop init.d breakage? 2014-09-25 12:06:39 i'll push new raspberry pi tarball after that... 2014-09-25 12:15:42 fabled: yes im on it with ncopa 2014-09-25 12:22:23 clandmeter: what do you think of this: http://sprunge.us/IQbD 2014-09-25 12:23:29 possibly add a check if modloop is empty and error out with "no modloop found" 2014-09-25 12:25:20 yes its fine from what i can tell. 2014-09-25 12:25:36 i think we can also stop look for .cmg 2014-09-25 12:29:04 also if you set modloop=/path/ as boot option, you probably want it to error out if it is not found 2014-09-25 12:29:19 rather than try find other 2014-09-25 12:33:57 http://www.bbc.com/news/technology-29361794 2014-09-25 12:34:01 or a warning 2014-09-25 12:34:01 did you guys see that 2014-09-25 12:34:33 we know about it, but i didnt read the bbc. 2014-09-25 12:34:38 oh you did :P 2014-09-25 12:34:38 is it on CNN as well now? 2014-09-25 12:34:41 ACTION sees patch 2014-09-25 12:34:45 idk 2014-09-25 13:30:15 <[xming]> BBC bashes bash 2014-09-25 13:33:46 clandmeter: http://sprunge.us/TXTT 2014-09-25 13:33:55 i got caught up in other stuff 2014-09-25 13:33:59 ugh, i forgot my coffe too 2014-09-25 13:37:57 looks ok 2014-09-25 14:14:30 ncopa: thx! 2014-09-25 21:15:36 N8@all 2014-09-26 01:27:31 welp 2014-09-26 01:27:34 openrc 2014-09-26 01:27:35 is dead 2014-09-26 01:31:28 ? 2014-09-26 03:05:42 kaniini: ?_? 2014-09-26 03:19:24 redhat hired one of the main guys i heard 2014-09-26 06:52:32 Guest97650: do you have link on redhat hiring openrc guy? 2014-09-26 06:55:12 <[xming]> hiring as in ditching systemd in favour of openrc or as burying openrc? 2014-09-26 06:56:18 <[xming]> does the buildbot builds pkgs with -Os CFLAG? 2014-09-26 06:56:27 yup 2014-09-26 07:02:11 <[xming]> what about some optimasation sensitve apps? 2014-09-26 07:03:06 <[xming]> like for minetest-server I like at least -O2 2014-09-26 07:09:53 [xming]: i suppose we can override CFLAGS where it makes sense 2014-09-26 07:10:00 i think we do so some places 2014-09-26 07:10:05 like zlib 2014-09-26 07:10:30 # we trade size for a little more speed. 2014-09-26 07:10:30 export CFLAGS="$CFLAGS -O2" 2014-09-26 07:34:30 <[xming]> ncopa: ah okay thanks 2014-09-26 12:40:59 ncopa: backuppc package is broken, i also noticed that there is no x (chmod) for others for dirs /etc /var/log and /var/lib, so backuppc user cannot do anything, it starts working after chmod +x to this dirs 2014-09-26 12:41:33 rnalrd: ^^^ its your pkg 2014-09-26 13:08:38 ncopa, i never used it, but I've packaged it for someone else 2014-09-26 13:08:44 never got such report 2014-09-26 14:54:17 hi 2014-09-26 14:54:29 is alpine secured against the bash bug? 2014-09-26 14:58:23 it has the first set of patches 2014-09-26 14:58:28 i dunno about the second set 2014-09-26 14:58:34 but bash isnt the default anyways 2014-09-26 14:58:45 ash (from busybox) is 2014-09-26 14:59:22 yeah, i am using zsh also 2014-09-26 14:59:28 but maybe they can call bash somehow 2014-09-26 15:01:16 cant call bash if it isnt installed 2014-09-26 15:01:18 :] 2014-09-26 15:11:16 nice, http://lwn.net/Articles/601840/ 2014-09-27 08:34:50 <[xming]> new patches for bash, the first one didn't solve the issue completely 2014-09-27 08:35:59 <[xming]> http://ftp.gnu.org/gnu/bash/bash-4.3-patches/bash43-026 2014-09-27 08:36:21 <[xming]> As far as I read, this is still not the final patch 2014-09-27 08:38:56 <[xming]> moin 2014-09-27 08:41:13 <[xming]> GNU bash, version 4.3.26(1)-release (x86_64-alpine-linux-musl) 2014-09-27 08:44:28 <[xming]> bash bump to 4.3.26 -> http://sprunge.us/XPTE 2014-09-27 10:34:27 <[xming]> it doesn't seem to fix CVE 2014-6271 for good, http://seclists.org/oss-sec/2014/q3/771 2014-09-27 11:28:41 please git pull git://git.alpinelinux.org/fab/aports 2014-09-28 12:44:21 does alpine linux have any plans of using a more simple and lean init system? 2014-09-28 12:44:35 such as runit, sinit, s6, etc. 2014-09-28 15:29:17 ncopa: ↑ 2014-09-29 06:23:37 morning 2014-09-29 06:23:52 fab 2014-09-29 06:24:04 >>> snowball: Checking sha512sums... 2014-09-29 06:24:04 libstemmer_c.tgz: FAILED 2014-09-29 06:24:04 sha512sum: WARNING: 1 of 1 computed checksums did NOT match 2014-09-29 06:24:45 the problem is that upstream just dumps new tarball with same name: http://snowball.tartarus.org/dist/ 2014-09-29 07:19:18 fab: i rebased it and removed the snowball update. snowball APKBUILD needs refactorung 2014-09-29 07:32:08 ncopa: thanks for pulling. i will put snowball on my todo list 2014-09-29 07:32:21 or have fcolista to do it 2014-09-29 07:33:43 https://github.com/snowballstem/snowball/issues/7 2014-09-29 08:35:41 ncopa: hi, about chromium, i'm stuck on incomplete type, could you give some advices please (p.s. i saw some patches for iproute2 with the same error) 2014-09-29 08:36:24 k0r10n: can i see your APKBUILD? 2014-09-29 08:36:28 sprunge < APKBUILD 2014-09-29 08:38:00 ncopa: http://sprunge.us/CNii 2014-09-29 08:38:16 but need to clean up... 2014-09-29 08:43:49 k0r10n: do you have apk build for ninja too? 2014-09-29 08:44:36 ncopa: http://sprunge.us/dgUT 2014-09-29 08:44:46 but also need to clean up :):) 2014-09-29 08:50:15 ncopa: i think you need some patches also, or you want just test ? 2014-09-29 08:51:57 i'm cleaning up ninja 2014-09-29 08:55:56 you build it with clang? 2014-09-29 08:56:16 yes 2014-09-29 08:56:28 why not our defautl compiler? 2014-09-29 08:57:26 tried to get it working, but without success 2014-09-29 09:29:47 it's strange, tested with default compiler and it worked , but same error 2014-09-29 09:52:51 hi 2014-09-29 09:53:10 re: snowball package in testing. 2014-09-29 09:53:37 There's not git repo available 2014-09-29 09:54:13 does abuild snapshot supports svn? 2014-09-29 09:54:26 afaik no 2014-09-29 09:57:22 i think you can use this: https://github.com/snowballstem/snowball 2014-09-29 09:59:11 looks that there are more that one snowball repo.. 2014-09-29 10:00:42 oh, that is a git repo 2014-09-29 10:00:50 8) 2014-09-29 10:23:19 uh-oh. 2014-09-29 10:23:34 sphinx doesn't compile 2014-09-29 10:23:56 there's a conflict between PAGE?SIZE const defined in /usr/include/limits.h 2014-09-29 10:24:06 and PAGE_SIZE defined in sphinx.cpp 2014-09-29 10:24:29 I think this came out with musl 2014-09-29 10:25:09 umh 2014-09-29 10:25:17 i can try to patch it 2014-09-29 10:25:37 there's a bug from 2011 from cygwin 2014-09-29 10:25:48 and is still open 2014-09-29 10:38:42 fcolista, while you are looking at sphinx could you update it to v2.2.4 pls ? 2014-09-29 10:41:36 vkrishn, this is actually what i'm doing 2014-09-29 10:41:44 wondering testing/xkcdpass should be named py-xkcdpass ? 2014-09-29 10:41:56 and this triggers the problem 2014-09-29 10:42:02 ok 2014-09-29 10:42:25 thanks, v2.2.4 much awaited feature version 2014-09-29 10:43:21 ok, i've a patched versionù 2014-09-29 10:43:23 version 2014-09-29 10:44:56 btw vkrishn 2014-09-29 10:45:02 since you are using it 2014-09-29 10:45:11 ok...vkrishn is gone.. 2014-09-29 10:47:37 still here, also checking logs :) 2014-09-29 10:47:46 so always around in `ghost mode` 2014-09-29 10:48:55 I checked sphinx/xapian in v2.7 and earlier AL, will try to do in v3.0 asap, was waiting for 2.2.4 2014-09-29 10:50:12 I think I would favor sphinx over xapian and once tested request to move to main/ 2014-09-29 11:54:00 vkrishn, i'm back. Sphinx compiled, now iàm going to push. Care to test it_ 2014-09-29 11:57:30 fcolista, ok would do, need to request clandmeter for v3.x setup 2014-09-29 11:58:43 vkrishn, to what 3.0 are you referring to? Alpine? 2014-09-29 11:59:00 yes 2014-09-29 11:59:11 This package is for Alpine 3 2014-09-29 12:00:38 so would need musl,edge setup 2014-09-29 12:01:10 yes 2014-09-29 12:01:36 upgraded packages are not backported generally 2014-09-29 12:02:00 re xkcdpass: fabian_a is the mantainer 2014-09-29 12:02:55 ok 2014-09-29 12:03:06 let me know if u need something else 2014-09-29 12:03:58 do you think I need to post a voting kinda on ML for xapian vs sphinx ? 2014-09-29 12:04:43 xapian also has some nice plans this year, http://trac.xapian.org/wiki/GSoCProjectIdeas 2014-09-29 12:05:21 why they cannot co-exists? 2014-09-29 12:05:40 I have no issue with it 2014-09-29 12:06:00 so I would re-test both and request to be pushed to main/ 2014-09-29 12:06:38 Ok. I don't see any reasons to support only one of the two (actually, we have more than one package who does the same) 2014-09-29 12:06:54 Thx for your tests. 2014-09-29 12:08:11 thanks for the updates 2014-09-29 12:08:24 np 2014-09-29 12:08:44 ncopa: re chromium and clang 2014-09-29 12:08:56 ncopa: resulting binary is a lot smaller without performance penalty 2014-09-29 12:09:07 ncopa: and it usually causes less pain 2014-09-29 12:13:22 ok 2014-09-29 12:13:39 seems like it downloads a toolchain 2014-09-29 12:13:41 binutils 2014-09-29 12:17:39 vkrishn: it seems that xkcdpass now provides a module. i think renaming would make sense 2014-09-29 12:22:55 ok, also there are pkgs for amqp, rabbitmq, any chance of having them in AL ? 2014-09-29 12:24:03 #1400 2014-09-29 12:28:29 ncopa, what happened to algibot, not picking hash-nums ? 2014-09-29 13:24:09 ncopa: about chromium, it seems it tries to use _libc_fpstate which is in ucontext.h in glibc distros, but same thing (but only _fpstate) is in bits/signal.h in AL 2014-09-29 16:16:57 somebody know where i can find execinfo.h in AL? 2014-09-29 16:17:09 this is from glibc headers 2014-09-29 16:26:30 there is no execinfo.h in musl 2014-09-29 16:26:57 this is non-standard header file 2014-09-29 16:27:57 you can try to redefine the provided functions by noop 2014-09-29 16:29:01 you're still working on chromium, aren't you? 2014-09-29 16:29:05 yes 2014-09-29 16:29:37 i found that i can define if i got musl, so it will avoid execinfo.h include 2014-09-29 16:30:05 long shot, but there is some possibility that chromium devs were sane enough to place backtrace functions inside some #ifdefs 2014-09-29 16:30:10 i can compile only 1876/15293 files :) 2014-09-29 16:30:21 but their check is too crappy to catch musl 2014-09-29 16:30:22 ouch 2014-09-29 16:30:25 that looks bad :) 2014-09-29 19:05:50 my packaging skills are so rusty… 2014-09-29 19:05:56 is there something like $subpkgname? 2014-09-29 20:22:24 fc1979, sphinx-2.2.4 not yet build on edge for x86 and x86_64 ? 2014-09-29 20:23:26 but seems to have for armhf 2014-09-30 07:25:46 failed to build v2.5.4-346-g8e27e51. http://distfiles.alpinelinux.org/distfiles/build-2-5-x86_64.txt 2014-09-30 08:39:14 failed to build v2.5.4-348-g3e3e698. http://distfiles.alpinelinux.org/distfiles/build-2-5-x86_64.txt 2014-09-30 08:41:24 build64-2-5: retry 2014-09-30 13:04:03 hello all, somebody know why malloc.h is empty ? 2014-09-30 13:04:50 k0r10n, all relevant stuff is defined in stdlib.h in the current musl. however, next musl release is going to have valid malloc.h 2014-09-30 13:05:09 i think malloc.h is legacy, and posix says malloc() is in stdlib.h 2014-09-30 13:05:12 fabled: when will be next release ? 2014-09-30 13:05:34 but as malloc_usable_size() was added that goes in malloc.h only, they added a real malloc.h 2014-09-30 13:05:44 k0r10n, edge has real malloc.h already 2014-09-30 13:06:00 fabled: ok, tyvm :) 2014-09-30 13:06:03 new musl release was due out mid-September. but it's not out yet. not sure what is the delay. 2014-09-30 13:06:20 we 3.0-stable gets it later when musl is tagged 2014-09-30 13:06:28 edge has musl git snapshot 2014-09-30 13:09:09 one more question while musl-wiki is not available, why there is no macro for musl ? 2014-09-30 13:09:20 fcolista, sphinx-2.2.4 not yet build on edge for x86 and x86_64 ? 2014-09-30 13:10:01 it builds 2014-09-30 13:10:13 vkrishn, yesterday i fixed it 2014-09-30 13:16:13 fcolista, still not build 2014-09-30 13:16:48 and not seeing more build servers in irc ?? 2014-09-30 13:19:02 amazingly some pkgs pushed after sphinx-2.2.4 were build yesterday, and could not locate the logs either 2014-09-30 13:19:51 vkrishn, sphinx-2.2.4 pushed yestarday built. Where is not building? 2014-09-30 13:20:00 In your build env? 2014-09-30 13:21:03 but teeworlds got build ;) 2014-09-30 13:33:11 ncopa: why doesnt sphinx get build? 2014-09-30 13:33:23 there is no log either. 2014-09-30 13:33:50 oh i think i know 2014-09-30 13:34:02 duh no i dont :) 2014-09-30 13:42:24 fcolista, I am checking http://nl.alpinelinux.org/alpine/edge/testing/x86/ for the pkgs 2014-09-30 13:42:32 use other ? 2014-09-30 13:47:35 dunno why it doesn't build actually vkrishn 2014-09-30 13:49:14 would be nice to have a raw logfile from builder, instead of only per package. 2014-09-30 13:51:28 ok now getting to my mqtt msgs 2014-09-30 13:55:17 wow , thats nice seeing new additions in msgs :) 2014-09-30 14:11:23 ncopa, thats weird, mqtt backlog msgs show commit id "9c2fb035c9b49" dated "Fri Sep 19" pushed after "Sat Sep 20" msgs 2014-09-30 14:11:46 bug in mqtt ? 2014-09-30 14:21:44 vkrishn: no, its a git feature 2014-09-30 14:32:11 ncopa: see my question? 2014-09-30 14:33:02 what arch? 2014-09-30 14:34:33 due to dependency i think 2014-09-30 14:34:36 snowball 2014-09-30 14:34:50 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64/testing/snowball/snowball-0_git20140929-r0.log 2014-09-30 14:41:59 ok, same day logs get overwritten 2014-09-30 14:44:55 oh, coz snowball pkg rel was not updated 2014-09-30 14:46:36 because it was a build fix 2014-09-30 14:46:46 ok 2014-09-30 14:47:01 pkgrel only needs update when the resulting pkg is modified 2014-09-30 14:47:09 but generally logs files are append mode 2014-09-30 14:47:27 ncopa: is there no general log where we can check whats going on? 2014-09-30 14:47:27 in this case there was no previous pkg 2014-09-30 14:47:58 http://bld2.alpinelinux.org/buildlogs/build-edge-x86_64.log and http://bld1.alpinelinux.org/buildlogs/build-edge-x86.log 2014-09-30 14:48:18 i have intention to provide a build.a.o web page with build server status 2014-09-30 14:48:20 and links to build logs 2014-09-30 14:48:23 but... 2014-09-30 14:48:51 the list of things i intend to do increases faster than i manage to actually do 2014-09-30 14:49:06 would it be ok if buildlogs could be fetched via rsync ? 2014-09-30 14:49:42 or just not useful 2014-09-30 14:52:58 vkrishn: i'd rather nfs mount it 2014-09-30 14:53:40 I mean could check what's changed by just getting the ls output 2014-09-30 14:54:07 giving quick overview 2014-09-30 15:05:12 good afternoon 2014-09-30 15:05:21 how are things in the alpine linux world? 2014-09-30 18:08:43 HI, I need some help getting started. I have a working chroot environment and abuild runs. The part I don't understand is how to build a complete bootable image which including the kernel bzImage (or equal) and initrd (or equal). Can someone point me in the right direction? Thanks 2014-09-30 18:16:21 Anyone there who can answer a question? 2014-09-30 18:23:01 sisaacs: You can use Makefile for it, see wiki 2014-09-30 18:23:37 http://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image 2014-09-30 18:23:59 thanks 2014-09-30 21:24:18 ncopa, do you guys have a patch for pth + musl 2014-09-30 22:03:43 I did an ISO build but it only created an ISO using prebuilt packages. I want to build everything from source. Is there a way to do that without having to do one package at a time?