2013-09-01 01:45:00 i am working on a glusterfs port 2013-09-01 01:49:24 /buffer 3 2013-09-01 01:49:25 oops 2013-09-01 02:53:45 could someone please update the prosody package when they have some time 2013-09-01 02:53:53 0.9.0 has been released 2013-09-01 23:53:27 grumble :P 2013-09-02 04:31:47 Moinmoin 2013-09-02 06:09:04 When I build a package with apkbuild, where does the apk go? 2013-09-02 06:09:22 abuild* 2013-09-02 06:10:30 morning 2013-09-02 06:10:46 Niichan: depends on what PKGDEST and/or REPODEST is set to 2013-09-02 06:10:51 in /etc/abuild.conf 2013-09-02 06:11:01 if both are unset then it goes to current dir 2013-09-02 06:13:17 ncopa: I see 2013-09-02 06:13:37 morning 2013-09-02 06:16:55 ncopa: Is there a way to install an arbitrary apk file a-la dpkg? 2013-09-02 06:20:01 apk add 2013-09-02 06:20:14 Sweet 2013-09-02 06:24:11 ncopa, i've realized that abuild install all the dependencies, but it doesn't uninstall them after a succesfull building. 2013-09-02 06:24:28 Something is strange with my env ? 2013-09-02 06:24:49 it creates .makedepends-$package with all the dependencies 2013-09-02 06:24:56 but it remains installed. 2013-09-02 06:25:25 This cause, when rebuild, that dependencies are met... 2013-09-02 06:25:36 and it can be misleading 2013-09-02 06:25:55 it should get removed 2013-09-02 06:26:01 yep 2013-09-02 06:26:13 at least it used to do it here 2013-09-02 06:26:55 barthalion, do you have cache dir in you apk/repositories ? 2013-09-02 06:27:02 haven't time to recreate containers yet 2013-09-02 06:27:07 No, default settings 2013-09-02 06:27:12 k 2013-09-02 06:27:15 like me 2013-09-02 06:27:16 just installed alpine-sdk, set some NOPASSWD in sudoers 2013-09-02 06:27:33 same here 2013-09-02 06:27:37 I'll check it after breakfast, but probably ncopa will confirm it by then as well 2013-09-02 06:28:17 i'm sure that it should be uninstalled after a building..ljust wondering why it does not happens :) 2013-09-02 06:28:46 fcolista: new feature in abuild 2013-09-02 06:28:51 aaaaaahh 2013-09-02 06:29:04 feature? :p 2013-09-02 06:29:21 add ERROR_CLEANUP="deps" to /etc/abuild.conf 2013-09-02 06:29:30 It's what i tought since it started after an upgrade :) 2013-09-02 06:29:38 and add CLEANUP="deps" 2013-09-02 06:30:03 CLEANUP="pkgdir srcdir deps" 2013-09-02 06:30:14 will remove the srcdir, the pkgdir and the deps on successful build 2013-09-02 06:30:16 neat 2013-09-02 06:30:27 I removed pkgdir and srcdir manually 2013-09-02 06:30:27 ERROR_CLEANUP sets what should be cleaned up on failed build 2013-09-02 06:30:49 should be documented :-/ 2013-09-02 06:30:56 just like vcs support :D 2013-09-02 06:31:05 Ok :) 2013-09-02 06:32:37 other pretty neat feature is: apk del .makedepends* 2013-09-02 06:33:38 hey barthalion, when you have time could you please upgrade the prosody package? 0.9 was finally released :P 2013-09-02 06:33:56 Mp5shooter: no problem, I should do it today 2013-09-02 06:34:05 awesome, thank you :) 2013-09-02 06:34:07 much appreciated 2013-09-02 06:34:11 alongside with Lua 5.2 support, do I remember correctly? 2013-09-02 06:34:12 its still in testing 2013-09-02 06:34:45 probably 2013-09-02 06:36:11 we aim for lua5.2 if possible right? 2013-09-02 06:36:44 hmm 2013-09-02 06:36:53 their install docs say 5.1 2013-09-02 06:37:18 asking them in their chat channel if 5.2 is good 2013-09-02 06:37:26 thanks 2013-09-02 06:37:30 np 2013-09-02 06:42:54 I asked if 0.9.0 supports lua 5.2 2013-09-02 06:42:56 "No. We are mostly compatible with Lua 5.2, but not entirely. Also, all Lua libraries need to be compiled with 5.2 headers." 2013-09-02 06:43:08 "Lua 5.2 can be considered a separate language. It's not compatible with 5.1." 2013-09-02 06:43:18 i know its not compatible 2013-09-02 06:43:31 and i know the libs needs to be 5.2 too 2013-09-02 06:43:38 ok 2013-09-02 06:43:43 oh ok 2013-09-02 06:43:44 so no official support for 5.2 2013-09-02 06:43:51 guess not 2013-09-02 06:44:02 they are working on 5.3 now... 2013-09-02 06:44:18 maintaing both 5.2 and 5.1 is heavy burden as it is 2013-09-02 06:44:37 i will not maintain lua5.1 and 5.3 at same time 2013-09-02 06:44:58 just ignore 5.3 2013-09-02 06:45:10 yeah, we'll deal with it when it comes out 2013-09-02 06:45:18 but when it does, i'd like drop support for 5.1 2013-09-02 06:45:18 5.2 isn't wide adopted yet 2013-09-02 06:45:32 Hmm, I seem to be having trouble with this APKBUILD: http://git.io/H3453g 2013-09-02 06:45:56 sounds like it's _almost_ fully compatible with 5.2 2013-09-02 06:45:58 "If you are interested in the code, the only thing left for full compatibility is getting rid of all the module("…") calls, which no longer work in 5.2." 2013-09-02 06:46:03 "Which is mostly a mechanical change. All the hard work is done." 2013-09-02 06:46:07 Output I get: https://gist.github.com/Niichan/7a126e695c403da124d1 2013-09-02 06:46:08 Niichan: i thought we alreay had ii? 2013-09-02 06:46:33 ncopa: You don't have ii 2013-09-02 06:46:48 oh... must have been since gentoo days then... 2013-09-02 06:46:56 run: file configure 2013-09-02 06:47:06 or check first line in configure 2013-09-02 06:47:15 i bet it says: #!/bin/bash 2013-09-02 06:47:21 if it does, add bash to makedepends 2013-09-02 06:47:32 Oh that's right 2013-09-02 06:47:38 II does not have a ./configure 2013-09-02 06:47:48 :) 2013-09-02 06:48:31 Mp5shooter: i think we wait for 0.9.1, with lua 5.2 2013-09-02 06:48:43 ok 2013-09-02 06:49:47 does prosody run as root? 2013-09-02 06:49:52 does it *need* to run as root? 2013-09-02 06:49:52 no 2013-09-02 06:50:00 i'm looking at the config pathc 2013-09-02 06:50:02 it runs as it's own user 2013-09-02 06:50:14 ncopa: How do you make patches again? 2013-09-02 06:50:25 ncopa: As in the patches that abuild will accept 2013-09-02 06:50:33 Niichan: -p1 2013-09-02 06:50:44 patch -p1 old_file new_file? 2013-09-02 06:50:50 thats how its applied 2013-09-02 06:50:58 I mean to create them 2013-09-02 06:51:08 so: cd src/foo-1.0/ 2013-09-02 06:51:24 diff -rupN orig.dir new.dir 2013-09-02 06:51:31 diff -u ./path/to/file.orig ./path/to/file > ../../foo.patch 2013-09-02 06:51:43 other way to do it: 2013-09-02 06:51:51 or init git repository 2013-09-02 06:52:03 abuild clean unpack; cd src/foo-1.0/; git init; git add .; git commit -minit 2013-09-02 06:52:07 2013-09-02 06:52:14 git diff > ../../foo.patch 2013-09-02 06:52:28 or even better, create a patch for upstream 2013-09-02 06:52:35 send it upstream 2013-09-02 06:52:38 wait for feedback 2013-09-02 06:52:58 once ACK'ed from upstream add it to aports 2013-09-02 06:53:33 ii prefers to install itself to /usr/local/ 2013-09-02 06:53:47 I'm just making a patch that changes config.mk to prefer /usr/ 2013-09-02 06:53:55 Niichan: you can most likely do: make PREFIX=/usr 2013-09-02 06:54:25 make install prefix=/usr DESTDIR="$pkgdir" 2013-09-02 06:55:03 \o/ 2013-09-02 06:55:09 Okay I get it now 2013-09-02 06:55:14 Thanks ncopa 2013-09-02 06:55:47 send to alpine-devel: http://wiki.alpinelinux.org/wiki/Creating_patches 2013-09-02 06:55:51 and i'll review it 2013-09-02 06:56:07 I will soom 2013-09-02 06:56:27 Just added it to my local repo 2013-09-02 06:56:58 commit message: 2013-09-02 06:57:05 testing/ii: new aport 2013-09-02 06:57:09 2013-09-02 06:57:22 Description what ii is 2013-09-02 06:57:25 2013-09-02 06:57:33 and thats it 2013-09-02 06:57:35 Will do 2013-09-02 06:58:31 something like this: http://git.alpinelinux.org/cgit/aports/commit/testing?id=5df815a4992c60cd0496e46c0b391cb2b5aca47a 2013-09-02 07:05:37 i'm tired, good night 2013-09-02 07:05:53 o/ 2013-09-02 07:05:58 \o 2013-09-02 07:06:27 ncopa: I tried to use git send-email and it failed 2013-09-02 07:06:48 do you have error message? 2013-09-02 07:07:09 ncopa: https://gist.github.com/Niichan/51fab7f9d3e7c97dee81 2013-09-02 07:07:42 http://wiki.alpinelinux.org/wiki/Development_using_git#Email_configuration 2013-09-02 07:07:57 setup sendemail.smtpserver 2013-09-02 07:08:50 ncopa: It should be in your inbox 2013-09-02 07:09:20 thanks 2013-09-02 07:09:22 ncopa: ERROR_CLEANUP=... / CLEANUP=... to /etc/abuild.conf or to each APKBUILD? 2013-09-02 07:09:34 ncopa: No problem :) 2013-09-02 07:09:47 StarWarsFan: abuild.conf 2013-09-02 07:10:21 correct 2013-09-02 07:10:34 so it's working for all packages 2013-09-02 07:10:35 ncopa: If it is okay with you guys I'm gonna be contributing more packages 2013-09-02 07:10:53 that is ofcourse very nice, thanks 2013-09-02 07:10:54 wouldn't it be nice to be able to use it on certain package builds? 2013-09-02 07:10:55 ncopa: Is it against the rules to contribute a package that builds from git? 2013-09-02 07:11:11 we usually create a snapshot and put it on our server 2013-09-02 07:11:20 Niichan: not against the rules, but we normally prefer releases 2013-09-02 07:11:29 because we have to maintain what we ship 2013-09-02 07:11:35 I understand 2013-09-02 07:11:41 ACTION will wait for cheat to get a release 2013-09-02 07:12:00 sometimes it helps to ask upstream 2013-09-02 07:12:26 It's a pretty new project 2013-09-02 07:12:28 https://github.com/chrisallenlane/cheat 2013-09-02 07:12:42 Gives you nice cheatsheets for *NIX commands 2013-09-02 07:13:25 206 commits 2013-09-02 07:13:25 EX: the one I threw in for apk: https://github.com/chrisallenlane/cheat/blob/master/cheatsheets/apk 2013-09-02 07:13:38 ask for a release 2013-09-02 07:14:03 nice project actually 2013-09-02 07:14:05 Just email him and ask for a release? 2013-09-02 07:14:06 too bad its python 2013-09-02 07:14:17 i'd create an "issue" in github 2013-09-02 07:14:55 ACTION nods 2013-09-02 07:15:19 cheat is a really nice idea 2013-09-02 07:15:25 should be trivial to implement it in C 2013-09-02 07:15:41 I should do that 2013-09-02 07:15:46 I need to get back to speed in C 2013-09-02 07:15:52 or even shell 2013-09-02 07:16:03 shell would be better imo 2013-09-02 07:16:09 yeah 2013-09-02 07:16:12 Which shell? 2013-09-02 07:16:17 posix /bin/sh... 2013-09-02 07:16:17 #!/bin/sh 2013-09-02 07:16:38 ACTION goes off to work on that 2013-09-02 07:16:48 Probably have to call it cheat-ng locally 2013-09-02 07:17:09 i'm working on gluster for 2.7 2013-09-02 07:17:11 and xen 4.3 2013-09-02 07:17:15 ncopa: actually… is it possible to remove *.la files automatically? 2013-09-02 07:17:19 never call a released product -ng ... 2013-09-02 07:17:20 (i have built 4.3 packages locally, we are testing) 2013-09-02 07:17:30 ncopa: it would be handy too 2013-09-02 07:17:35 barthalion: not really, i do it manually 2013-09-02 07:17:48 ncopa: I mean in my local ~/code/ 2013-09-02 07:18:01 I'll take a look at the codebase later 2013-09-02 07:18:22 Niichan: thats is ofcourse good, just to release it as a product ... :) 2013-09-02 07:18:43 ncopa: I gotta do the git repo dance 2013-09-02 07:18:46 barthalion: the reason i dont do it is because i want the apkbuilds be more transparent 2013-09-02 07:19:06 I have this APKBUILD for cheat already: https://github.com/Niichan/within-aports/blob/master/testing/cheat/APKBUILD 2013-09-02 07:19:16 barthalion: i want the apkbuild show whats going on 2013-09-02 07:19:20 https://github.com/lucaswerkmeister/cheats 2013-09-02 07:19:27 this might work already on /bin/sh 2013-09-02 07:19:51 find "$pkgdir" -name '*.la' -delete || return 1 2013-09-02 07:19:54 reminds me 2013-09-02 07:19:55 sh: Doesn't work at all. Even if you fix the easy stuff (function x -> x(), [[ ... ]] -> [ ... ]), it still does some weird stuff so that the filenames are messed up. Not sure why. 2013-09-02 07:22:40 ACTION goes to remove redundant code in cheats 2013-09-02 07:24:06 oh, oops 2013-09-02 07:25:36 ncopa: How quickly do the mirrors update? 2013-09-02 07:25:48 depends on the mirror admin 2013-09-02 07:25:53 normally an hour or 2 2013-09-02 07:25:57 lax-noc? 2013-09-02 07:26:01 dunno 2013-09-02 07:26:36 ncopa: ah, I see 2013-09-02 07:27:20 kaniini: It seems the guy that made cheats (the shell script) really beat around the bush in making it. 2013-09-02 07:27:46 i'd like to see a port of 9base to alpine btw 2013-09-02 07:27:52 rc shell would be really nice to have 2013-09-02 07:27:59 I should probably get shadowircd in alpine >_> 2013-09-02 07:59:19 ncopa: Reject that patch I just made 2013-09-02 08:00:05 ok? 2013-09-02 08:00:11 ncopa: I will fix it and re-submit 2013-09-02 08:12:00 ncopa: It's all good now 2013-09-02 08:24:04 Niichan: i have a few nitpicks on shadowircd 2013-09-02 08:24:13 ncopa: Oh? 2013-09-02 08:24:20 just trivial stuff 2013-09-02 08:24:32 ncopa: Is it the having to call it /usr/bin/ircd stuff? 2013-09-02 08:24:32 i will only fix the whitespace damages 2013-09-02 08:24:37 no 2013-09-02 08:24:42 just trivial 2013-09-02 08:24:49 like whitespace at end of line 2013-09-02 08:25:05 should be pkgrel=0 2013-09-02 08:25:06 etc 2013-09-02 08:25:08 Oh, I'll try to avoid that in the future 2013-09-02 08:25:17 does not matter so i'll let it be as it is 2013-09-02 08:25:30 but i'll comment for the future 2013-09-02 08:25:42 there were are few more 2013-09-02 08:25:52 Niichan: do you use Vim? 2013-09-02 08:25:58 barthalion: Yes 2013-09-02 08:26:11 nnoremap W :%s/\s\+$//:let @/='' 2013-09-02 08:26:23 In the vimrc? 2013-09-02 08:26:24 W will remove all trailing whitespace from a file 2013-09-02 08:26:33 What is the leader key? 2013-09-02 08:26:44 not sure what's the default, probably \ 2013-09-02 08:27:05 my vimrc: http://sprunge.us/dQcY 2013-09-02 08:27:25 the 3 first lines will make whitespace at EOL red 2013-09-02 08:28:06 ncopa: ooooh, I see. Thanks :) 2013-09-02 08:32:16 fabled: hpn-ssh patch is in the news 2013-09-02 08:32:21 http://www.phoronix.com/scan.php?page=news_item&px=MTQ0OTQ 2013-09-02 08:32:28 i wonder *why* its news though... 2013-09-02 08:32:41 well, I didn't know about it 2013-09-02 08:32:47 so it was a news for me 2013-09-02 08:32:59 it's been in alpine for years 2013-09-02 08:33:16 yeah, same with gentoo 2013-09-02 08:33:45 fabled have used to forward port the patches to new openssh releases early 2013-09-02 08:38:38 actually 2013-09-02 08:38:41 we dont use it 2013-09-02 08:38:52 we only use the peakput 2013-09-02 08:39:08 which is for the progress meter 2013-09-02 08:39:35 and then we have something that looks to be for hw accel 2013-09-02 08:39:39 a one lines 2013-09-02 08:39:41 liner* 2013-09-02 08:41:13 fabled: do you think we should apply the dynwindows patch too? 2013-09-02 08:41:20 they are available for openssh 6.2 2013-09-02 08:46:47 ncopa, i thought we had dynwindows there 2013-09-02 08:46:54 someone removed them on early upgrade? 2013-09-02 08:46:58 and forgot to readd? 2013-09-02 08:47:09 likely 2013-09-02 08:47:15 i'll add it 2013-09-02 08:47:51 heh 2013-09-02 08:48:00 git blame points on barthalion :) 2013-09-02 08:48:11 0ba8a696 2013-09-02 08:49:49 oops! 2013-09-02 08:50:06 not a bit deal 2013-09-02 08:50:15 it just explains why scp has gone slow :) 2013-09-02 08:50:30 it wasn't applying, and I sucked too much then to port it 2013-09-02 08:51:24 I'll add a comment where to find the HPN patch upstream 2013-09-02 10:43:19 by the way 2013-09-02 10:43:23 i am thinking about 2013-09-02 10:43:31 setting up a "try alpine free for 1 hour" thing 2013-09-02 10:43:39 kinda like how ubuntu has it 2013-09-02 10:50:32 ncopa, et al.: what do you think regarding that 2013-09-02 10:51:09 personally, i think it could be a useful marketing tool 2013-09-02 10:55:48 ++ 2013-09-02 10:56:36 i think for people to really 'get' what makes alpine a good distribution, they have to actually sit down and use it for a minute 2013-09-02 11:02:23 kaniini: any news regarding build of debug-packages? 2013-09-02 11:02:44 i still have the issue with the crash of quasselcore... 2013-09-02 11:06:55 kaniini: i'm not familiar with the 1 hour free ubuntu thingy 2013-09-02 11:07:04 is it server hosting for 1 hour? 2013-09-02 11:07:11 so you can play around with the distro? 2013-09-02 11:07:26 yap 2013-09-02 11:07:33 sounds useful to me 2013-09-02 11:40:14 kaniini: do you nkow which init.d script that should create /var/run/xen? #2031 2013-09-02 11:40:42 my guess is that it is xendomains 2013-09-02 11:40:48 but i'm ust guessing 2013-09-02 11:48:40 ncopa: I think it should be xenstore 2013-09-02 11:49:18 ncopa: xendomains init script is optional IMHO, while /var/run/xen should always be there, even if the user decides to not use xendomains 2013-09-02 11:49:33 start-stop-daemon --start --exec /usr/sbin/xenstored \ 2013-09-02 11:49:33 --pidfile /var/run/xenstored.pid \ 2013-09-02 11:49:33 -- --pid-file=/var/run/xenstored.pid \ 2013-09-02 11:49:50 looks like xenstored uses /var/run and not /var/run/xen/ 2013-09-02 11:50:34 not for xenstore itself, but xenstore should always be started, while xendomains is optional 2013-09-02 11:52:37 ncopa: xensotre should do it 2013-09-02 11:52:52 xend.initd shouldnt really be used right? 2013-09-02 11:52:54 ok 2013-09-02 11:53:11 ncopa: not really, unless you plan to use xm (xend) 2013-09-02 11:53:49 ncopa: I guess most distros create it on xencommons, but we don't have that one 2013-09-02 11:54:33 it looks to me that all .initd scripts uses /var/run/.pid 2013-09-02 11:54:45 except xendomains 2013-09-02 11:55:16 so my intuition tells me it should be in xendomains... 2013-09-02 11:55:27 but i'll put it in xenstore 2013-09-02 11:56:19 I'm not sure if xl/qemu also use /var/run/xen/ (I think so) 2013-09-02 11:56:57 dev32-edge:~/aports/main/xen$ grep run xenqemu.initd 2013-09-02 11:56:57 #!/sbin/runscript 2013-09-02 11:56:57 --pidfile=/var/run/xenqemu.pid -b -- \ 2013-09-02 11:56:57 -pidfile /var/run/xenqemu.pid \ 2013-09-02 11:56:57 --pidfile /var/run/xenqemu.pid 2013-09-02 11:57:02 might be xl uses it though 2013-09-02 11:57:26 let me do a quick grep 2013-09-02 11:58:28 i couldnt find it with grep 2013-09-02 11:58:45 ncopa: yes, it is needed by libxl for runtime, so it should be created unconditionally (even if xendomains is not used) 2013-09-02 11:59:05 so xenstore 2013-09-02 11:59:33 yes, I think it is the best place 2013-09-02 11:59:43 or the least worst place 2013-09-02 12:16:40 rebooting brb 2013-09-02 12:25:06 fabled: how do i find out what is holding back libvirtd-1.1.2? https://dpaste.de/btXT4/ 2013-09-02 12:35:50 wierd, trying to force upgrade it does not generate any error: https://dpaste.de/Gbw8v/ 2013-09-02 12:36:10 the libvirt package has some install_ifs 2013-09-02 12:36:16 might be related.. 2013-09-02 12:38:16 kaniini: you've seen my question around 1pm? 2013-09-02 12:46:41 StarWarsFan: have not had time to look at it yet. 2013-09-02 12:48:54 ok 2013-09-02 12:48:58 another question: 2013-09-02 12:49:04 how to solve issues like this: 2013-09-02 12:49:16 apk upgrade 2013-09-02 12:49:16 ERROR: unsatisfiable constraints: 2013-09-02 12:49:16 libpng-1.6.3-r0: 2013-09-02 12:49:16 conflicts: libpng-1.5.17-r0 2013-09-02 12:49:16 satisfies: freetype-2.5.0.1-r1[so:libpng16.so.16] 2013-09-02 12:49:17 libpng-1.5.17-r0: 2013-09-02 12:49:19 conflicts: libpng-1.6.3-r0 2013-09-02 12:49:22 satisfies: openjdk6-jre-1.6.0_p24-r8[so:libpng15.so.15] 2013-09-02 12:50:08 try apk upgrade --available 2013-09-02 12:50:17 ncopa, not all libpng dependencies are recompiled 2013-09-02 12:50:33 err... not all packages depending on libpng 2013-09-02 12:50:33 kaniini: same error 2013-09-02 12:50:48 StarWarsFan, libpng got upgraded, but not everything is recompiled 2013-09-02 12:50:54 satisfies: openjdk6-jre-1.6.0_p24-r8[so:libpng15.so.15] 2013-09-02 12:50:57 x86_64? 2013-09-02 12:51:04 yes, edge 2013-09-02 12:51:10 ncopa, i have same issue on x86 2013-09-02 12:51:19 multiple packages depending on libpng15.so.15 2013-09-02 12:51:23 and can't ugprade 2013-09-02 12:51:27 fabled: you likely have packages from testing 2013-09-02 12:51:39 x86_64 is tricky becoase openjdk6 does not build atm 2013-09-02 12:51:42 i think there's both. some from testing, but some from main too 2013-09-02 12:52:12 i belive i updated all libpng15 deps in main, except openjdk6 on x86_64 2013-09-02 12:53:11 looks that cpan.org is experiencing problems 2013-09-02 12:53:32 fabled: could you please try find out which pkg that is missing in main/x86? 2013-09-02 13:10:06 hm, is there something changed with the handling of ${pkgdir} on APKBUILD? 2013-09-02 13:10:30 i have a build script, with this content: 2013-09-02 13:10:39 ... 2013-09-02 13:10:40 build() { 2013-09-02 13:10:41 mkdir -p ${pkgdir}/var/install/bin 2013-09-02 13:10:42 ... 2013-09-02 13:10:54 this was working all the time 2013-09-02 13:11:03 on edge- and 2.6-buildnodes 2013-09-02 13:11:21 but since some days this is failing on the edge-buildnodes 2013-09-02 13:12:06 seems like ${pkgdir}/var/install/bin is not created!? 2013-09-02 13:13:33 any ideas? 2013-09-02 13:16:55 a bit of bikeshedding… why would you create such directory? 2013-09-02 13:17:31 yeah, i wonder the same, why /var/.../bin ? 2013-09-02 13:18:11 there are some make steps afterwards which use this directory 2013-09-02 13:21:26 fcolista: if you have the tarball locally, could you please scp it to dev.alpinelinux.org:/var/cache/distfiles? 2013-09-02 13:26:34 yes ncopa 2013-09-02 13:27:14 weird, appears to only affect crul? 2013-09-02 13:27:16 curl* 2013-09-02 13:27:21 bah 2013-09-02 13:27:27 sometimes works, sometimes not 2013-09-02 13:27:31 it also happens with firefox 2013-09-02 13:27:35 varnish timeout 2013-09-02 13:27:42 ok.. 2013-09-02 13:27:51 i had same issue here with mod_perl 2013-09-02 13:27:57 but i was able to dl it with ff 2013-09-02 13:28:29 Error 503 Service Unavailable 2013-09-02 13:28:34 Service Unavailable 2013-09-02 13:28:34 Guru Meditation: 2013-09-02 13:28:35 XID: 1007593803 2013-09-02 13:28:35 Varnish cache server 2013-09-02 13:28:40 with ff 2013-09-02 13:28:42 http://www.cpan.org/authors/id/M/MS/MSCHOUT/ 2013-09-02 13:28:55 yeah 2013-09-02 13:28:59 they have issues with the backend 2013-09-02 13:29:00 after some reload it works 2013-09-02 13:29:04 yep 2013-09-02 13:29:12 do you still want me upload tar.gz ? 2013-09-02 13:29:37 yes please 2013-09-02 13:30:05 ok 2013-09-02 13:30:17 scp: /var/cache/distfiles/Apache-AuthCookie-3.19.tar.gz: Permission denied 2013-09-02 13:30:18 :) 2013-09-02 13:30:33 scp Apache-AuthCookie-3.19.tar.gz fcolista@dev.alpinelinux.org:/var/cache/distfiles 2013-09-02 13:32:27 ncopa, since i've no rights to copy on /var/cache/distfiles, i've copied to /home/fcolista 2013-09-02 13:33:43 copied 2013-09-02 13:33:44 thanks 2013-09-02 13:34:14 thx 2013-09-02 13:36:02 perl dependencies are a nightmare 2013-09-02 13:37:30 yeah.. 2013-09-02 13:37:43 apkbuild-cpan makes it easier 2013-09-02 13:37:53 yes...do you still have problem with mod_perl ? 2013-09-02 13:37:58 yes 2013-09-02 13:38:09 i have not been able to update mod_perl to 2.0.8 2013-09-02 13:38:09 I'm working on netdot right now, and apache2-mod-perl doesn't work 2013-09-02 13:38:37 yes mod_perl does not build 2013-09-02 13:38:39 unno why 2013-09-02 13:47:11 fabled: i cannot find out which package in main that is linked to libpng15 2013-09-02 13:47:43 gst-plugins-bad 2013-09-02 13:47:46 wxgtk 2013-09-02 13:47:50 openjpeg 2013-09-02 13:47:53 cairo 2013-09-02 13:47:57 qt-x11 2013-09-02 13:48:02 ghostscript 2013-09-02 13:48:07 hum 2013-09-02 13:48:12 try apk update? 2013-09-02 13:48:17 maybe some of them are wrong versions selected by logic 2013-09-02 13:48:29 libpng-1.5.17-r0: 2013-09-02 13:48:29 conflicts: libpng-1.6.3-r0 2013-09-02 13:48:30 breaks: libpng-dev-1.6.3-r0[libpng=1.6.3-r0] 2013-09-02 13:48:30 satisfies: openjpeg-1.5.1-r0[so:libpng15.so.15] 2013-09-02 13:48:30 wxgtk-2.8.12.1-r0[so:libpng15.so.15] 2013-09-02 13:48:30 cairo-1.12.14-r0[so:libpng15.so.15] 2013-09-02 13:48:34 qt-x11-4.8.4-r3[so:libpng15.so.15] 2013-09-02 13:48:36 gdk-pixbuf-2.28.2-r0[so:libpng15.so.15] 2013-09-02 13:48:38 freetype-2.5.0.1-r0[so:libpng15.so.15] 2013-09-02 13:48:40 that's with --latest 2013-09-02 13:48:43 openjpeg is all i think 2013-09-02 13:50:41 fabled: openjdk6 is a problem though 2013-09-02 13:50:48 it fails to build on x86_64 2013-09-02 13:50:55 i suspect it is since gcc-4.8 2013-09-02 13:52:23 what's the library "nsl"? 2013-09-02 13:52:24 No library found for -lnsl 2013-09-02 13:52:30 mod_perl still has issues with apache 2.4x 2013-09-02 13:52:37 for 2.2x is ok 2013-09-02 13:52:52 ncopa: ^^^ 2013-09-02 13:53:00 i know.. i just dont know how to fix it 2013-09-02 13:53:11 there is a 2.0.8 release available 2013-09-02 13:53:29 fcolista: i think its resolver lib 2013-09-02 13:53:35 which uclibc has no support for at all 2013-09-02 13:53:41 ah 2013-09-02 13:53:46 ok. 2013-09-02 13:53:50 iirc, 2.0.8 is still targeted for 2.2 2013-09-02 13:53:51 I have to skip the test then 2013-09-02 13:54:05 yeah 2013-09-02 13:54:07 remove l-nsl 2013-09-02 13:54:27 it's wone of the test made by make test perl script 2013-09-02 13:55:05 vkrishn: please file a bug for mod_perl... 2013-09-02 13:59:47 ncopa, will do, but I have not done any testing with it recently 2013-09-02 14:17:03 hi 2013-09-02 14:20:30 xl uses /var/run/xen 2013-09-02 14:22:49 ncopa: http://irclogger.com/.alpine-linux/2013-08-30#1377894690 2013-09-02 14:23:18 Could you also please check the forum, two new posts 2013-09-02 14:23:18 probably because nobody asked for it? 2013-09-02 14:24:10 eh no 2013-09-02 14:24:20 its the default 2013-09-02 14:24:24 compiled into kernel 2013-09-02 14:24:28 rather than module 2013-09-02 14:24:46 Ok 2013-09-02 14:25:14 i am glad he asked 2013-09-02 14:25:22 we should probably check if it could be built as module 2013-09-02 14:37:29 ncopa: can you take a look at poppler? we're behind some releases, but I remeber that I had issues with building 2013-09-02 14:39:17 ok 2013-09-02 14:39:32 there was some issue with poppler-gt being gtk3 only? 2013-09-02 14:39:43 i think there was some app that depends on gtk2 2013-09-02 14:39:54 oh, maybe that's it 2013-09-02 14:40:07 probably some old gnome apps 2013-09-02 14:40:22 we should upgrade evince to gtk3 2013-09-02 14:41:13 I'll take a look at some other updates now 2013-09-02 14:42:26 Do you have found the correct init file for /var/run/xen ? 2013-09-02 15:03:26 shafire: yea i pushed the fix already 2013-09-02 15:03:33 i hope it works :) 2013-09-02 15:04:19 uh, right 2013-09-02 15:04:23 I'm brainless 2013-09-02 15:06:11 I shouldn't rely on my script to automatize building yet… 2013-09-02 15:06:50 ncopa: can you update the bug with your commit and close it? 2013-09-02 15:10:17 what is more frustrating than perl dependencies are perl dependencies with cpan who does not work 2013-09-02 15:10:39 possibly yes 2013-09-02 15:10:51 broken DNS entries are also funny 2013-09-02 15:11:27 algitbot: better build this time 2013-09-02 15:13:24 ähm, to go back to the asked problem using $pkgdir on abuild: 2013-09-02 15:13:33 is there something changed on edge-abuild? 2013-09-02 15:13:56 no idea, it should work 2013-09-02 15:14:36 yea, "should" ;-) 2013-09-02 15:16:08 I guess that install -d doesn't work either? :) 2013-09-02 15:17:10 what do you mean? 2013-09-02 15:18:22 try it instead of mkdir 2013-09-02 15:20:45 ah, nice idea 2013-09-02 15:20:50 thx for the hint, will check... 2013-09-02 15:24:17 algitbot: good for you 2013-09-02 15:42:40 hm, seems like the $pkg* variables are not expanded 2013-09-02 15:43:02 another no longer working example: 2013-09-02 15:43:05 mkdir -p ${startdir}/pkg/${pkgname} 2013-09-02 15:43:37 the directory ".../pkg/" is created 2013-09-02 15:43:59 in which function do you run those? 2013-09-02 15:44:07 but not ".../pkg/$pkgname" :-( 2013-09-02 15:44:15 huh 2013-09-02 15:44:21 prepare 2013-09-02 15:45:12 well, $pkgdir shouldn't be available from prepare anyway 2013-09-02 15:45:41 was working all the time and is still working on 2.6-builders 2013-09-02 15:46:20 why not? 2013-09-02 15:46:38 because prepare is meant for work on sources 2013-09-02 15:47:42 i c 2013-09-02 17:10:44 <@ncopa> the 3 first lines will make whitespace at EOL red 2013-09-02 17:10:50 nice.. i hate trailing spaces 2013-09-02 17:34:46 hm 2013-09-02 17:34:54 what key is 2013-09-02 17:35:23 disables arrows 2013-09-02 17:35:51 oh 2013-09-02 17:36:30 i just installed vim to test out the red whitespace 2013-09-02 17:36:33 and wow 2013-09-02 17:36:38 the syntax coloring is so pretty 2013-09-02 17:39:02 no more vi lolz 2013-09-02 17:41:12 uh, vi 2013-09-02 18:20:51 ncopa: I submitted a fixed tor package 2013-09-02 23:45:51 What package is apkbuild-pypi in? 2013-09-02 23:46:22 Never mind, found it 2013-09-03 00:43:14 o.O 2013-09-03 00:43:18 what's the tor package called Niichan 2013-09-03 00:43:21 i didn't think anyone ever packaged it 2013-09-03 00:44:39 Mp5shooter: It's called not merged yet 2013-09-03 00:44:50 Mp5shooter: But if you want it naaaaaaaaao 2013-09-03 00:44:56 o 2013-09-03 00:44:59 Mp5shooter: I could email it to you or something 2013-09-03 00:45:06 when you said 'fixed' tor package i thought there was already one 2013-09-03 00:45:15 I flubbed submission of it ^^;; 2013-09-03 00:45:22 oh lol 2013-09-03 00:47:21 Mp5shooter: https://github.com/Niichan/within-aports/tree/master/testing/tor 2013-09-03 00:48:15 cool 2013-09-03 00:48:19 are you going to package torsocks too ;))) 2013-09-03 00:48:49 gag 2013-09-03 00:49:18 wat 2013-09-03 00:50:15 Idunno 2013-09-03 00:50:30 I packaged onioncat in my aports too 2013-09-03 00:51:19 A lot of them pull from git repos 2013-09-03 01:53:39 hmm 2013-09-03 01:53:58 is there a way to say that a certain package can't be installed if x package is installed? 2013-09-03 01:54:10 apk add !nano 2013-09-03 01:55:06 i mean like 2013-09-03 01:55:13 is there a way to put it in an APKBUILD 2013-09-03 01:55:53 so if i have lua-sec installed and i was trying to apk add lua-sec-specialversion it would stop and tell me i need to uninstall lua-sec first or something 2013-09-03 01:57:10 I was just thinking that would be useful for a shadowircd-git package I wanted to make 2013-09-03 01:57:40 heh 2013-09-03 01:57:46 yeah it would be useful for prosodyu 2013-09-03 01:57:52 it has its own version of lua-sec 2013-09-03 05:43:34 Mp5shooter: there is conflicts iirc 2013-09-03 05:44:03 or not 2013-09-03 05:44:33 well, at least not on wiki 2013-09-03 05:44:55 Mp5shooter: but I'm sure that abuild from git has conflicts variable 2013-09-03 06:00:10 morning 2013-09-03 06:00:17 today is 2.6.4 release day 2013-09-03 06:00:27 yay 2013-09-03 06:00:51 maintenance release, nothing exciting 2013-09-03 06:54:51 re 2013-09-03 06:54:52 moin 2013-09-03 06:55:21 morning 2013-09-03 07:49:15 ncopa: do you have an idea regarding the problem with $pkgname a/o $pkgdir on APKBUILD on edge-machines? 2013-09-03 07:49:43 what problem? 2013-09-03 07:49:51 see 17:13 yesterday 2013-09-03 07:50:06 you have backlog? 2013-09-03 07:51:08 hm 2013-09-03 07:51:24 i think the $pkgdir variable is expanded 2013-09-03 07:51:47 but with new release of abuild, it will remove pkgdir before running package() 2013-09-03 07:52:13 ? 2013-09-03 07:52:35 before calling package() { ... } it will rm -r "$pkgdir" 2013-09-03 07:52:44 or something like that 2013-09-03 07:53:01 uh... 2013-09-03 07:53:28 so you cannot create anything in "$pkgdir" before 'package' stage 2013-09-03 07:53:40 i c 2013-09-03 07:53:47 do you have an APKBUILD? 2013-09-03 07:54:06 sure 2013-09-03 07:54:08 i might give a suggestion how to adjust it 2013-09-03 07:54:18 just a second... 2013-09-03 07:58:35 can't remember: what was the cool pastebin-replacement alpine uses via cmd line? 2013-09-03 07:59:01 sprunge 2013-09-03 07:59:09 ahhh, thx 2013-09-03 07:59:18 need to keep that in mind... ;-) 2013-09-03 08:01:53 so, see here: 2013-09-03 08:01:54 http://sprunge.us/eZEN 2013-09-03 08:04:00 aha, you store the source in aports git itself 2013-09-03 08:05:10 you should do the install -D -m644 ${_builddir}/edit-conf.cui/ecelib ${pkgdir}/var/install/include/ecelib 2013-09-03 08:05:18 in package() too 2013-09-03 08:05:50 also, runtime binaries into /var is a bit against FHS 2013-09-03 08:06:00 maybe unless the code is selfmodifying 2013-09-03 08:13:15 ah ok 2013-09-03 08:13:27 will try this, thx for the hint 2013-09-03 09:55:09 whoops 2013-09-03 09:55:23 maybe should have had linux-virt-grsec updated in stable too? 2013-09-03 09:55:37 currently is only in edge 2013-09-03 09:55:51 i think is due to be included in 2.7 2013-09-03 09:55:53 ok 2013-09-03 09:57:07 rnalrd: i'm testing owncloud. http://wiki.alpinelinux.org/wiki/OwnCloud 2013-09-03 09:57:13 nice 2013-09-03 09:57:35 i'm not too happy for that chown -R that needs to be done everytime package is updated 2013-09-03 09:57:41 i think we need a www-data group 2013-09-03 09:57:46 and add webservers to that group 2013-09-03 09:57:47 yeah 2013-09-03 09:58:38 the /usr/share/webapps/owncloud/apps should probably be moved to /var/lib/owncloud/apps and symlinked 2013-09-03 09:58:50 that way we have no need for writing anything in /usr 2013-09-03 09:59:27 that was on my todo list 2013-09-03 09:59:32 agree 2013-09-03 09:59:41 i can file a bug 2013-09-03 09:59:43 tnx 2013-09-03 18:30:27 You missed a dependency on wpa_supplicant 2013-09-03 18:30:33 I just emailed in a patch to fix that 2013-09-03 20:01:24 n8 2013-09-03 21:35:28 "You asked about 5.2 support - it's partly done, but it probably won't be in the next release, possibly the one after" 2013-09-03 21:35:29 :( 2013-09-04 05:11:37 talking to libreswan devs, they say that with openrc, "start on already started should NEVER fail" 2013-09-04 05:11:41 and that it's in LSB? 2013-09-04 05:19:26 nvm 2013-09-04 05:26:31 turns out 2013-09-04 05:26:38 the dummy init script for libreswan was the issue :P 2013-09-04 05:26:47 looks like libreswan works with theirs 2013-09-04 07:54:16 hi 2013-09-04 07:54:28 .4 is available :) 2013-09-04 09:07:59 Mp5shooter: alpine does not follow LSB. 2013-09-04 11:59:51 bah 2013-09-04 11:59:59 apache2-mod-perl won't work 2013-09-04 12:00:09 apr_log undefined 2013-09-04 12:00:24 also ldd od mod_perl.so returns "no symbols" 2013-09-04 12:01:51 also a lot of undef libs 2013-09-04 12:01:52 http://pastebin.com/hWxe5hMK 2013-09-04 12:04:05 umh 2013-09-04 21:58:00 hm 2013-09-04 21:58:16 if something generates a custom init script when you build the source 2013-09-04 21:58:32 how do you get it into the APKBUILD that it should be used as the init script and placed in /etc/init.d 2013-09-04 22:38:38 Would it be a bad idea to run an ircd on grsec? 2013-09-04 23:58:50 Is there a way to verify that a patch is being applied in an APKBUILD file? 2013-09-04 23:59:00 Preferably some form of message like 2013-09-04 23:59:05 "FOO.patch applied" 2013-09-05 00:02:27 Ah, never mind 2013-09-05 00:02:35 I didn't add the patch to the list of sourcefiles 2013-09-05 00:32:07 When is zfs going to be available again? It is sort of a bummer to not be able to mount /home anymore. 2013-09-05 02:43:38 Niichan: I've run ircd on grsec without issues 2013-09-05 02:43:51 Niichan: both ngircd and charybdis 2013-09-05 02:43:59 they both seem to work pretty well 2013-09-05 02:45:42 why does pidgin not support irc on alpine? 2013-09-05 02:46:15 sorry, wrong channel 2013-09-05 02:47:28 ncopa: do you have an arm build environment yet? 2013-09-05 05:57:38 superduck: fabled has an arm build box. i dont have any arm box yet 2013-09-05 05:57:42 for development 2013-09-05 09:54:57 2.6.4 x86_64 need an "options usb_storage delay_use=15" usb times out 2013-09-05 11:26:18 hi 2013-09-05 11:26:19 hm 2013-09-05 11:26:38 there is a boot option usb_delay=15 2013-09-05 13:31:36 ok. i got my new arm bootstrap working. now to just make sata disk+ nic work, and we can setup builder on the wandboard. 2013-09-05 13:32:08 fabled, I've got some beaglebone blacks that can be used for building/testing 2013-09-05 14:06:50 superduck, cool. i got a wandboard quad here with 2GB ram. almost everything needed works. only SATA is left. after that can start builder setup. 2013-09-05 14:08:10 fabled, alright, I think my boards have 512MB of ram and 2GB of eMMC 2013-09-05 14:08:23 I can attach some USB storage or sshfs mount extra storage if needed 2013-09-05 14:09:30 fabled, right now I'm running kali linux on the beaglebone black, but I'd like to switch to alpine. 2013-09-06 11:08:22 urgh 2013-09-06 11:08:34 oops 2013-09-06 11:08:39 ncopa, linux-vanilla changes sneaked in 2013-09-06 11:08:44 wasn'tsupposed topush that 2013-09-06 11:09:00 and went along inthe sudo commit 2013-09-06 11:09:02 duh 2013-09-06 11:10:50 happens 2013-09-06 11:10:58 did it break things? 2013-09-06 11:11:12 you can do a partial revert if it did 2013-09-06 12:03:05 ncopa: anything berkely db 6.x ? 2013-09-06 12:03:17 not time for that this week 2013-09-06 12:03:20 why do you need it? 2013-09-06 12:03:28 is it backwards compat with db-5? 2013-09-06 12:03:53 I think they did license change 2013-09-06 12:04:09 but need to check back-compat 2013-09-06 12:04:20 is it abi compat? 2013-09-06 12:04:34 if its abi compat then we can just bump it 2013-09-06 12:05:15 would suggest db6 instead if in doubt 2013-09-06 12:05:22 in testing 2013-09-06 12:05:46 when tested could be moved main and replace current 5.x 2013-09-06 12:06:13 has mongodb been tested ? 2013-09-06 12:06:35 read in logs about it being large and has some issues 2013-09-06 12:06:45 Found 2 nice nosqls - riak, rethinkdb 2013-09-06 12:06:49 just in case 2013-09-06 12:11:04 looks like riak is only for IA-32, x86-64 2013-09-06 12:11:37 about db-6 I could not download it, they need registration..logging...etc 2013-09-06 12:12:21 any possibility to get src from dev.AL ? 2013-09-06 12:12:43 i dunno 2013-09-06 12:12:54 if we cannot simply wget it then we have problems 2013-09-06 12:13:04 or if we are not allowed to redistribute the sources 2013-09-06 12:14:39 change of license to AGPL 2013-09-06 12:52:22 i'm gonna stop build-edge for a while 2013-09-06 12:52:35 while trying to bootstrap ada 2013-09-06 13:12:45 i will stop build64-edge too 2013-09-06 13:12:53 and then i will push a gcc update that enables ada 2013-09-06 13:13:06 i have build ada and will install manually 2013-09-06 13:15:48 http://upstream-tracker.org/versions/libdb.html 2013-09-06 13:15:50 between 6.0.19 5.3.21 - bc ~ 47% and lots of abi changes 2013-09-06 13:15:50 and I am able to download now, current db-6.0.20.tar.gz 2013-09-06 13:16:10 so db6 ? 2013-09-06 13:16:27 maybe this could be added github.com/lvc/abi-compliance-checker.git 2013-09-06 13:17:00 ~ 42% 2013-09-06 13:19:00 now download complete... checking changelogs 2013-09-06 13:46:25 ok here comes ada... 2013-09-06 13:47:30 nice 2013-09-06 13:47:47 vkrishn: ignore db6 2013-09-06 13:48:11 vkrishn: license has changed from MIT to AGPLv6 2013-09-06 13:48:39 due to that you cannot link other programs with non compatible license to db6 2013-09-06 13:49:27 likely are commercial apps 2013-09-06 13:54:40 no, it's only compatible with GPL3 and enforces AGPL terms 2013-09-06 13:55:04 Debian had this issue already and we (as in Arch Linux) dropped the db6 rebuild month ago 2013-09-06 13:57:29 so what happens to db depend apps ? 2013-09-06 13:57:29 mdb? 2013-09-06 13:59:06 you need to use some alternative or drop non-compatible apps 2013-09-06 13:59:31 very often you can disable db at all or just tdb (etc), but sometimes it's not possible 2013-09-06 14:02:34 grep -rle 'db-dev' ./* 2013-09-06 14:02:42 i'm somewhat interested in an Alpine/kFreeBSD 2013-09-06 14:02:47 would anyone be interested in that 2013-09-06 14:02:51 = 12+12 main/testing 2013-09-06 14:03:33 vkrishn: and for example, transition from postfix built with db to postfix without it requires manual intervention 2013-09-06 14:11:11 I am getting to like mediagoblin.org :) 2013-09-06 14:14:11 kaniini: sounds interesting, but I'm barely skilled in anything, so I doubt that I can help 2013-09-06 14:15:18 btw, there's taking bdb from freebsd and packaging that up 2013-09-06 14:15:21 same api 2013-09-06 18:17:35 barthalion: Is it possible to build a mosh package on Alpine? Useful app, but uClibc may be causing problems. 2013-09-07 10:14:21 hm 2013-09-07 10:14:38 i'm trying to repackage libreswan so it installs its own init script 2013-09-07 10:15:15 when I do abuild -r I later get an error that says something along the lines of "cannot create regular file /etc/init.d/ipsec: permission denied" and the build fails 2013-09-07 10:15:17 o.O 2013-09-07 16:35:45 nvm 2013-09-07 16:35:50 tried again this morning and it worked 2013-09-07 16:35:51 weird 2013-09-07 16:35:51 :P 2013-09-07 17:08:16 hm 2013-09-07 17:08:20 perl uses /usr/local 2013-09-07 17:08:23 is it supposed to 2013-09-07 17:08:25 ? 2013-09-07 20:33:40 hm 2013-09-07 20:33:58 I had to comment the export LDFLAGS line in abuild.conf to get luasec-prosody to build 2013-09-07 21:23:31 ok i made a patch which should fix it 2013-09-07 21:23:52 it won't patch during abuild though.. 2013-09-07 21:23:59 idk what i'm doing wrong 2013-09-07 21:28:00 could you paste your APKBUILD somewhere? 2013-09-07 21:28:06 sure 2013-09-07 21:28:40 http://s.mp5.im/f0k9 2013-09-07 21:29:25 I usually apply patch manually 2013-09-07 21:29:59 I mean, without for loop 2013-09-07 21:29:59 oh 2013-09-07 21:30:27 less handy, but you gain more control 2013-09-07 21:31:37 http://s.mp5.im/4tg6 2013-09-07 21:31:39 that's what happens :( 2013-09-07 21:32:15 algitbot: build master 2013-09-07 21:32:32 ah, wrong strip level 2013-09-07 21:32:36 try -p0 2013-09-07 21:32:41 oh 2013-09-07 21:32:42 lol 2013-09-07 21:32:45 I tried p1 etc 2013-09-07 21:32:49 didn't know there was a p0 :P 2013-09-07 21:33:05 yeah, used when there is no directory to strip 2013-09-07 21:33:19 you probably should leave -p1 but change your patch 2013-09-07 21:33:25 ---- a/Makefile 2013-09-07 21:33:31 ++++ b/Makefile 2013-09-07 21:33:33 oh ok 2013-09-07 21:33:34 or something like that 2013-09-07 21:33:43 it's always some silly mistake with me ;P 2013-09-07 21:34:08 nobody is perfect :p 2013-09-07 21:36:20 hmm 2013-09-07 21:36:24 http://s.mp5.im/Uph5 2013-09-07 21:36:28 does this mean it's a bad patch 2013-09-07 21:36:49 yup 2013-09-07 21:36:56 hm 2013-09-07 21:37:30 I go to bed 2013-09-07 21:37:34 good night 2013-09-07 21:38:08 good night 2013-09-07 21:38:20 and thanks :P 2013-09-07 21:57:15 oh.. 2013-09-07 21:57:21 it's trying to patch the wrong makefile 2013-09-07 21:57:21 wow 2013-09-07 21:57:29 just noticed that 2013-09-07 22:27:52 well 2013-09-07 22:27:54 it's a good patch 2013-09-07 22:28:02 I put in the full path when it asks for it and it worsk 2013-09-07 22:28:03 works* 2013-09-07 22:28:13 just can't get the patch settings right somewhere ;? 2013-09-07 22:28:15 :/ 2013-09-07 22:45:33 AHA 2013-09-07 22:45:33 GOT IT 2013-09-07 23:11:41 git: 'sendmail' is not a git command. See 'git --help'. 2013-09-07 23:11:42 o.O 2013-09-07 23:11:44 er 2013-09-07 23:11:48 that was one of my tests 2013-09-07 23:11:52 send-email isn't working 2013-09-07 23:27:43 oh 2013-09-07 23:27:45 had to install git-perl 2013-09-08 04:41:02 hmm 2013-09-08 11:30:44 please git pull git://git.alpinelinux.org/fab/aports 2013-09-08 11:31:01 there are a couple of updates and some new aports (incl. deps) 2013-09-08 12:28:10 god, almost 150 commits 2013-09-08 12:28:42 what about poppler, does it work for you? 2013-09-08 12:30:33 yay freenode lag 2013-09-08 12:31:53 fabian_a: I'll reword some commits to present tense 2013-09-08 12:37:34 fabian_a: I won't push poppler, consult with ncopa about it 2013-09-08 12:41:10 I bet I will regret not building it locally before push 2013-09-08 12:52:30 barthalion: thanks 2013-09-08 12:54:26 fabian_a: I'll just revert troublesome commits, feel free to push them again 2013-09-08 12:54:45 barthalion: i will talk with ncopa about poppler 2013-09-08 12:58:57 barthalion: i will fix py-barcodenumber as soon my build system is back to normal. 2013-09-08 12:59:10 sure, no problem 2013-09-08 12:59:37 barthalion: thanks again for taking care of my commits 2013-09-08 13:04:55 fabian_a: no, I thank you for your changes :p 2013-09-08 13:15:51 i think some of those ruby commits will break redmine. 2013-09-08 13:16:15 drop me a line which, we should revert them either 2013-09-08 13:16:29 keep it 2013-09-08 13:17:37 i will create redmine ruby pacakges later, when i got time. 2013-09-08 13:17:57 ruby is a b*itch to maintain properly 2013-09-08 13:18:12 oops didnt remove the i :) 2013-09-08 13:18:18 haha :D 2013-09-08 13:21:47 2 week of italy with lots of wine and limoncello does that to the brain. 2013-09-08 13:31:17 xen-4.3 hitting momentarily. 2013-09-08 13:35:50 hmm! 2013-09-08 13:36:01 seriously? 2013-09-08 13:36:08 algitbot: retry build64-edge 2013-09-08 13:36:24 algitbot: retry master 2013-09-08 13:36:28 algitbot: build master 2013-09-08 13:36:30 oh 2013-09-08 13:36:32 okay 2013-09-08 13:51:51 something bugs me though 2013-09-08 13:52:11 "coroutine backend ucontext" 2013-09-08 13:52:13 on qemu 2013-09-08 14:19:42 oh, right. 2013-09-08 14:19:51 xen-hypervisor isn't available on 32-bit anymore 2013-09-08 15:21:40 http://kaniini.dereferenced.org/alpine/2013/09/08/advanced-abuild-hacking.html 2013-09-08 16:29:58 I submitted a couple of patches to alpine-devel, could someone please put em in 2013-09-08 16:30:15 (hope I emailed em right) 2013-09-09 08:54:13 algitbot: build master 2013-09-09 08:59:04 algitbot: build master 2013-09-09 09:03:53 algitbot: build master 2013-09-09 09:04:39 hello arm build server 2013-09-09 09:04:54 you are more than welcome :) 2013-09-09 09:39:17 algitbot: i love you 2013-09-09 09:40:18 fabled: did i get back just in time? 2013-09-09 09:48:31 clandmeter, ? 2013-09-09 09:51:27 fabled: arm 2013-09-09 09:51:31 oh 2013-09-09 09:51:32 right 2013-09-09 09:51:39 i guess. it's work-in-progress. 2013-09-09 09:51:51 the board works ok? 2013-09-09 09:51:54 yes 2013-09-09 09:52:03 better then expected? 2013-09-09 09:52:08 i have created initial bootstrap; and have alpine running on wand board currently 2013-09-09 09:52:16 builder is mostly setup 2013-09-09 09:52:29 but i need to do quite a bit of musl fixes to aports still 2013-09-09 09:52:33 i saw you added more cross support. thts just to bootstrap? 2013-09-09 09:52:37 yes 2013-09-09 09:52:49 i use now abuild to create cross-compiler too 2013-09-09 09:53:11 i have quite large binutils+gcc aports changes pending 2013-09-09 09:54:07 fabled: the first rpi images will still take some weeks? 2013-09-09 09:54:16 depends on how things go 2013-09-09 09:55:34 can probably do some limited images early. but for full alpine aports, it's gonna take quite a while. 2013-09-09 09:56:22 i'm now redoing the bootstrap with ada enabled 2013-09-09 09:56:31 then will do native gcc self-build 2013-09-09 09:56:37 and then start fixing aports 2013-09-09 10:13:31 hi 2013-09-09 10:15:21 how are you guys? :) 2013-09-09 10:15:40 hi 2013-09-09 10:15:41 im bad 2013-09-09 10:15:48 im back in rain and cold 2013-09-09 10:16:28 here it's sunny :) 2013-09-09 10:16:39 but had an examen some hours ago :S 2013-09-09 10:24:18 bootstrapping ada to arm/musl now 2013-09-09 10:24:25 seems it needs minor patching 2013-09-09 10:25:45 thanks for doing it 2013-09-09 10:29:28 ok. cross compiler built. now to see if the native compiler builds with it. 2013-09-09 10:39:45 how long does it need for a roun? 2013-09-09 10:39:48 how long does it need for a round? 2013-09-09 10:43:30 not too long 2013-09-09 10:43:31 failed 2013-09-09 10:43:47 home/tteras/alpine-sysroot/armv6-alpine-linux-muslgnueabihf/usr/include/sched.h:112:36: error: attempt to use poisoned "calloc" 2013-09-09 10:43:47 #define CPU_ALLOC(n) ((cpu_set_t *)calloc(1,CPU_ALLOC_SIZE(n))) 2013-09-09 10:48:31 http://gcc.gnu.org/ml/gcc-patches/2004-09/msg02421.html 2013-09-09 10:48:44 is using pragma a solution? 2013-09-09 10:49:31 i think it's including stuff in wrong order 2013-09-09 10:49:35 should be simple to fix 2013-09-09 11:02:17 shafire: we have ada now in the default x86/x86_64 toolchain 2013-09-09 11:02:21 apk add gcc-gnat 2013-09-09 11:02:38 ncopa: I saw it, thanks. 2013-09-09 11:02:52 thanks for your bootstrap toolchain 2013-09-09 11:02:57 ncopa: Do I need to watch out for something? 2013-09-09 11:03:11 no i think not 2013-09-09 11:03:26 i thin it will just work 2013-09-09 11:11:48 no way to see apache2-mod-perl work 2013-09-09 11:13:03 you will have to port it to apache 2.4 2013-09-09 11:13:06 well 2013-09-09 11:13:10 yes.. 2013-09-09 11:13:12 i think it might work fro current svn 2013-09-09 11:13:31 i tryed the last one svn commit (3 days ago) but still won't compile 2013-09-09 11:15:16 ok, i think i figured out the ada issue 2013-09-09 11:20:08 What was the problem? 2013-09-09 11:21:15 ncopa: http://alpinelinux.org/apk/main/x86/gcc-gnat <- from where did you copy paste the description? 2013-09-09 11:23:08 http://pkgs.fedoraproject.org/cgit/gcc.git/tree/gcc.spec#n603 2013-09-09 11:24:37 shafire, basically bad #include order 2013-09-09 11:24:48 but now there's another problem related to cross compiling 2013-09-09 11:28:34 ncopa: Ah, okay. gcc 4.8 is Ada 2012 (or at least 2005) already 2013-09-09 11:28:47 ncopa: gcc 4.9 will be ada 2012 by default 2013-09-09 12:52:04 jbilyk_: hi, did you find the issue reg backups on www? 2013-09-09 12:53:08 clandmeter: ncopa and I talked about it, but I didn't fix it yet 2013-09-09 12:53:14 sorry...haven't had time 2013-09-09 12:59:28 hmm ok 2013-09-09 14:09:45 ncopa: that 0mq buildmsg-pub/sub is still active ? 2013-09-09 14:12:31 was reading about alert-project.eu and it reminded me that we have something similar ongoing in AL dev 2013-09-09 14:13:36 we we still use buildmsg-pub/sub 2013-09-09 14:15:53 any plans to add other sensors 2013-09-09 14:17:42 "other sensors?" 2013-09-09 14:18:14 there is a pdf "ALERT Conceptual Architecture.pdf" 2013-09-09 14:18:22 moment would try to get the link 2013-09-09 14:19:51 http://www.alert-project.eu/sites/portal2-alert.atosorigin.es/files/content-files/download/ALERT%20Conceptual%20Architecture.pdf 2013-09-09 14:20:05 from this page http://www.alert-project.eu/content/alert-conceptual-architecture-0 2013-09-09 14:21:06 what is alert-project? 2013-09-09 14:22:08 more like "Knowledge extractor for un/structured information" for a development process 2013-09-09 14:22:54 different sensors(event based) for BTS/SCM...etc that feed the process 2013-09-09 14:23:18 sensors work on already existing infra 2013-09-09 14:25:09 here pub/sub uses tomcat/jetty.... 2013-09-09 14:25:34 finally. ada native compiler cross-built 2013-09-09 14:25:57 cross build fails due to http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55946 2013-09-09 14:26:01 the one interface you build on 0mq is much simpler 2013-09-09 14:26:59 see page 33 of the pdf link 2013-09-09 14:32:11 gotta go, do leave msg if any, would be reading logs 2013-09-09 17:37:14 hi 2013-09-09 17:37:32 fabled: I read that the ada compiler was successfully built? 2013-09-09 17:38:00 yup, see gcc-gnat package 2013-09-09 17:39:20 barthalion: I talking about the arm platform :) 2013-09-09 17:39:32 I am talking about the arm platform :) 2013-09-09 17:40:02 ah 2013-09-09 20:15:47 1546 MiB in 388 packages 2013-09-09 20:15:55 Is that Megabytes or mibibits? 2013-09-09 20:20:26 mebibytes 2013-09-09 20:21:07 How does that convert to megabytes? 2013-09-09 20:23:15 mega stands for 10^6 2013-09-09 20:23:22 mebi is 2^20 2013-09-09 21:43:48 yeah ncopa you were right 2013-09-09 21:43:49 :P 2013-09-09 21:43:53 libreswan is all wonky 2013-09-09 21:43:58 with that init script 2013-09-09 21:44:01 sigh 2013-09-09 23:25:12 What version is Alpine 2.5's GCC? 2013-09-10 01:25:50 files from v2.5.4-188-g742b9d6 uploaded 2013-09-10 01:25:58 files from v2.5.4-188-g742b9d6 uploaded 2013-09-10 01:29:15 files from v2.4.11-90-g5da0603 uploaded 2013-09-10 01:29:45 files from v2.3.6-265-g82f5fcb uploaded 2013-09-10 02:09:43 does anyone know what package has pkcs11-tool? 2013-09-10 02:23:55 ncopa, kaniini: can someone with commit access add this patch to testing/atheme-services? http://git.io/zHJzHA I would email it in but git is being a pain 2013-09-10 02:35:05 that patch is incomplete 2013-09-10 02:35:08 checksums etc. 2013-09-10 02:36:07 Lemme redo it 2013-09-10 02:36:44 Ah, was ssh'd into the wrong box 2013-09-10 02:47:29 kaniini: https://gist.github.com/Heartmender/d8379b92a6996333d9b0 2013-09-10 03:12:50 ugh 2013-09-10 03:12:54 easy-rsa isn't being easy 2013-09-10 03:13:25 keeps wanting me to "source ./vars" which doesn't seem to work right 2013-09-10 03:15:36 maybe if i switch my shell to /bin/sh 2013-09-10 03:15:37 hm 2013-09-10 03:16:56 damn 2013-09-10 03:16:58 didn't help 2013-09-10 03:18:26 ugh 2013-09-10 03:18:43 ok whatever, easy-rsa 2013-09-10 03:18:46 i'll do it by hand 2013-09-10 03:47:19 AHA 2013-09-10 03:47:28 replaced `'s with double quotes 2013-09-10 07:24:20 hi 2013-09-10 07:25:57 ncopa: you have now a own ada bootstrap, or? 2013-09-10 07:32:03 shafire: yes we have it bootstrapped already 2013-09-10 07:32:17 ah 2013-09-10 07:32:18 yes 2013-09-10 07:32:27 we have crosscompile too 2013-09-10 07:32:36 for bootstrapping ada on musl/arm 2013-09-10 07:32:58 I read it yesterday:) cool 2013-09-10 08:26:04 shafire, yes, i got ada for arm cross-compiled 2013-09-10 08:29:57 so, it should also work, when we switch to x86/musl, or? 2013-09-10 08:34:00 shafire, yes 2013-09-10 08:34:34 :-) 2013-09-10 08:36:00 i also patched few things. e.g. ada to produce dynamic linked binaries by default 2013-09-10 08:37:05 again: thanks for doing it :) 2013-09-10 08:42:14 fabled: when will you have the first alpha release? 2013-09-10 08:42:29 shafire, of what? 2013-09-10 08:42:41 alpine on arm 2013-09-10 08:43:16 dunno yet when to publish. i already have bootable image for wand board. 2013-09-10 09:56:36 ncopa, i think gcc fixes need new abuild 2013-09-10 12:51:57 algitbot: build master 2013-09-10 13:20:58 Who wrote apk? 2013-09-10 13:21:16 shafire, me and ncopa 2013-09-10 13:24:36 in C, or? 2013-09-10 13:25:22 shafire, the version 1.x was shell script collection. the current 2.x+ version is C 2013-09-10 13:25:43 ah okay. 2013-09-10 13:26:07 ncopa wrote the original stuff. and i rewrote it in C with various optimizations. and now it's been we fixing and fine tuning it together. 2013-09-10 13:32:48 Do you know any anti virus program other than clamav, which will run on alpine? 2013-09-10 13:33:17 I tested f-prot, the scanner works, only the signature-updater is linked against glibc :( 2013-09-10 13:35:03 shafire: no antivirus on alpine 2013-09-10 13:35:10 except clamav 2013-09-10 13:35:19 there are not many opensource antiviruses 2013-09-10 13:36:06 imho, antivirus is more a way to make money on than solving a problem 2013-09-10 13:37:02 fabled: do we need the PKGDEST vs mluti arch for subpackages support? 2013-09-10 13:37:12 before i tag a new abuild release 2013-09-10 13:37:16 so i can push new gcc 2013-09-10 13:37:53 imo, having subpackages that have different arch is a can of worms 2013-09-10 13:45:59 ncopa, not must. we need to do it eventually. but for now, it's ok. 2013-09-10 13:48:46 ok 2013-09-10 13:49:00 hm 2013-09-10 13:49:05 but we need new abuild for gcc, right? 2013-09-10 13:49:18 i think we could workaround it in APKBUILD 2013-09-10 13:49:32 just set CTARGET_{ARCH,LIBC} to C{ARCH,LIBC} if they are not defined 2013-09-10 13:49:45 but i'd like to have new abuild regardless 2013-09-10 13:51:09 i'll attempt to do native musl/arm build of gcc now 2013-09-10 14:03:44 I'll tag an alpine-edge snapshot 2013-09-10 14:03:45 on request 2013-09-10 15:04:36 ncopa: new iso fails to boot in VirtualBox 2013-09-10 15:04:47 edge iso? 2013-09-10 15:04:55 yes 2013-09-10 15:05:17 whats the error message? 2013-09-10 15:06:09 i need to go 2013-09-10 15:06:17 i will test it tm in qemu/kvm 2013-09-10 15:07:36 sha1 passes 2013-09-10 15:08:11 Booting kernel failed: Invalid argument 2013-09-10 23:01:04 hm 2013-09-10 23:01:12 were there any recent kernel changes in edge? 2013-09-11 05:56:57 Mp5shooter: i upgraded kernel to 3.10.11 yes 2013-09-11 14:00:40 pplibvirtneeds a bump 2013-09-11 14:00:40 oops 2013-09-11 14:00:40 libvirtneeds a bump 2013-09-11 14:00:40 libvirt needs a bump 2013-09-11 14:00:40 that's what is causing Dave1029's problem 2013-09-11 14:47:40 oh thanks! 2013-09-11 14:47:43 i'll bump it 2013-09-11 14:54:31 hm 2013-09-11 14:54:38 i dont know what goes wrong... 2013-09-11 14:54:51 ERROR: unsatisfiable constraints: 2013-09-11 14:54:51 Huh? Error reporter did not find the broken constraints. 2013-09-11 14:57:05 kaniini: ERROR: xen-4.3.0-r1: trying to overwrite usr/libexec/qemu-bridge-helper owned by qemu-1.6.0-r1. 2013-09-11 14:57:24 hmm 2013-09-11 14:57:27 okay 2013-09-11 14:57:45 ncopa: how do you propose we handle that conflict 2013-09-11 14:58:08 hm 2013-09-11 14:58:31 maybe --libexecdir=/usr/lib/xen and --libexecdir=/usr/lib/qemu 2013-09-11 14:59:31 but i think it is something else that breaks the build64-edge 2013-09-11 14:59:35 i dunno what 2013-09-11 15:01:14 ok i have to go 2013-09-11 15:46:44 kaniini: i disabled libvirt for now. 2013-09-11 15:47:14 ncopa: ^ 2013-09-11 16:25:12 ncopa: okey dokey. i will investigate it. 2013-09-11 18:35:49 hi 2013-09-11 19:59:11 hi 2013-09-11 19:59:20 I want to build a mono apk 2013-09-11 19:59:29 should I use it with llvm or not? 2013-09-12 06:18:28 morning 2013-09-12 06:18:48 re libvirt, i cannot make it fail here locally 2013-09-12 06:19:00 it could be an apk-tools bug? 2013-09-12 06:19:04 mornin 2013-09-12 06:19:27 i tried that init script for libreswan that you emailed 2013-09-12 06:19:32 didn't work 2013-09-12 06:19:32 :( 2013-09-12 06:24:09 actually 2013-09-12 06:24:10 hm 2013-09-12 06:24:20 it was written of the top of my head 2013-09-12 06:24:25 it most likely needs minor fixes 2013-09-12 06:24:43 it might have typos etc 2013-09-12 06:38:30 quick question 2013-09-12 06:38:34 if something is looking for /proc/sys/crypto/fips_enabled 2013-09-12 06:38:41 what package would have what it needs? 2013-09-12 06:38:44 (FIPS support) 2013-09-12 06:40:32 sounds like a kernel module 2013-09-12 06:41:26 oh 2013-09-12 06:43:53 https://bugzilla.redhat.com/show_bug.cgi?id=786145 2013-09-12 06:44:06 looks like we need disable CRYPTO_MANAGER_DISABLE_TESTS 2013-09-12 06:44:44 oh 2013-09-12 06:44:51 I was just asking because libreswan can operate in FIPS mode 2013-09-12 06:44:52 :P 2013-09-12 06:44:58 ACTION hasn't come across that sshd bug 2013-09-12 06:46:04 fabled: do you mind have a look at this? 2013-09-12 06:46:14 do we want CONFIG_CRYPTO_FIPS? 2013-09-12 06:46:22 This options enables the fips boot option which is 2013-09-12 06:46:22 required if you want to system to operate in a FIPS 200 2013-09-12 06:46:22 certification. You should say no unless you know what 2013-09-12 06:46:23 this is. 2013-09-12 06:46:29 i dotn know what it is... :) 2013-09-12 06:49:39 http://en.wikipedia.org/wiki/Federal_Information_Processing_Standards 2013-09-12 06:49:49 appears to be a US government standard 2013-09-12 06:50:20 "The purpose of FIPS is to ensure that all federal government and agencies adhere to the same guidelines regarding security and communication." 2013-09-12 06:50:45 so Mp5shooter, unless you work for the US gov, why bother about FIPS? 2013-09-12 06:51:15 idk :P 2013-09-12 06:53:08 algitbot: build master 2013-09-12 07:59:35 Hi all, I have a problem with bgp and zebra ... Zebra does not inject any routes ( static or from bgp ) could this be an sec kernel issue ??? 2013-09-12 08:11:11 hi 2013-09-12 08:11:22 Is it allowed to use git in APKBUILD? 2013-09-12 08:11:51 not really preferred 2013-09-12 08:18:51 mh, okay. mono-llvm 3 can only be build from git :( 2013-09-12 08:30:54 if upstream dont want make a release (that they maintain), why should we? 2013-09-12 08:33:21 what is this ?? anybody ?? openvpn ... write UDPv4: Invalid argument (code=22) 2013-09-12 08:41:57 fossil has a new version 2013-09-12 13:08:47 hmm - cloning aports seems pretty slow today ~300K/s 2013-09-12 13:08:52 dropping as low as 150 2013-09-12 14:47:55 ncopa: did you see the latest edge iso returns an error with setup-alpine? it does boot though. 2013-09-12 14:49:34 syslinux trigger returns error 1 2013-09-12 14:51:04 and apk always returns an error 2013-09-12 14:51:06 Executing uclibc-utils-0.9.33.2-r23.trigger 2013-09-12 14:51:06 1 errors; 206 MiB in 39 packages 2013-09-12 15:02:13 no ididnt see that 2013-09-12 15:02:16 will check on monday 2013-09-12 15:02:20 will be out most of the day tm 2013-09-12 15:29:23 ooops ... problem with shorewall 4.5.20 ... once it's started it messes with kernel routes and zebra cannot inject routes anymore ... ideas ??? 2013-09-12 15:42:49 argg 2013-09-12 15:43:24 looks like some ruby packages are not bumped after ruby version bump. 2013-09-12 15:43:42 ruby2? 2013-09-12 15:43:53 i think so 2013-09-12 15:43:58 1.9 to 2.0 2013-09-12 15:46:53 mh, isn't there a ruby package manager, so the packages are not needed? 2013-09-12 15:47:20 sure there is a package manager 2013-09-12 15:47:24 but we have our own 2013-09-12 15:47:46 like npm for nodejs 2013-09-12 15:47:56 there is nothing easier then apk add redmine (and it just works) 2013-09-12 15:48:34 let me see how big the damage is 2013-09-12 15:49:05 why not something like apk add ruby, gem install redmine (and it just works :)) 2013-09-12 15:50:31 wow, there a lot ruby packages 2013-09-12 15:51:36 I think, many of these packages could be installed easily with gem too 2013-09-12 15:52:07 http://alpinelinux.org/apk/main/x86_64/ruby-fcgi vs http://rubygems.org/gems/fcgi 2013-09-12 15:52:46 "This version aims to be compatible with both 1.8.x and 1.9.x versions of Ruby, and also will be ported to 2.0.x." 2013-09-12 15:55:03 we have gem and its installed by default 2013-09-12 16:17:15 atleast main is ok now 2013-09-12 16:29:48 okay 2013-09-12 16:29:53 see you later! 2013-09-12 23:58:12 ugh 2013-09-12 23:58:18 screwing with the ipsec init script 2013-09-12 23:58:31 i somehow made pluto go into an infinite loop of dying and restarting itself 2013-09-12 23:58:54 it dies because grsec keeps killing it for some unknown annoying reason 2013-09-12 23:59:21 why can't anything be easy :( 2013-09-13 10:43:02 ncopa: seems ruby pkg is broken 2013-09-13 10:43:42 its splitted into subpkg's but gem is listing them as installed making it spit out errors . 2013-09-13 11:05:23 clandmeter, you around ? 2013-09-13 11:05:33 sure 2013-09-13 11:06:15 I was thinking if gem packages were handled by both apk and gem together, this could solve some problem 2013-09-13 11:06:59 explain 2013-09-13 11:07:18 what apk add would do is it would drop the gem package in cache dir eg. ruby/lib/ruby/gems/1.9.1/cache 2013-09-13 11:07:29 then invoke gem install 2013-09-13 11:07:44 that way both apps are happy 2013-09-13 11:08:13 only problem would be taking care of index file relating to gem app 2013-09-13 11:08:46 to me that seems like more hassle 2013-09-13 11:08:55 from end-user pov 2013-09-13 11:09:17 like ? 2013-09-13 11:10:04 if apk has correct ruby packages, one could just apk add ruby-something and it would work 2013-09-13 11:10:42 yes, apk add would do the same, as said above 2013-09-13 11:11:06 only that it would have a post.install script to invoke gem 2013-09-13 11:11:18 but we are already doing that 2013-09-13 11:11:25 but on the buildserver 2013-09-13 11:11:34 we invoke gem to install the gems 2013-09-13 11:11:50 i dont understand what the adv is to do it post install 2013-09-13 11:12:33 imho the way apk deals with gems is ok. there is only one issue 2013-09-13 11:12:35 well 2 atm 2013-09-13 11:13:12 our current ruby base is corrupt 2013-09-13 11:13:21 ah 2013-09-13 11:13:38 our ruby package makes subpkgs of a few base gems 2013-09-13 11:14:10 so it does not install them by default. but gems *does* think its installed. 2013-09-13 11:14:46 so we can do 2 things: make it *not* see the missing gems, or just install them when we install ruby. 2013-09-13 11:14:55 making "subpkgs of base gem" - will that not confuse gem app ? 2013-09-13 11:15:13 I dont know how to do that first option, but i do know how to fix the second option. 2013-09-13 11:15:33 In alpline we try to strip as much as possible to make it modular. 2013-09-13 11:15:56 but its has its downsides sometimes... like now. 2013-09-13 11:16:32 but upgrade path from 1.9 -> 2.0 was not optiomal (to say the least). 2013-09-13 11:18:20 so I first rebuild all ruby pkgs against 2.0 to fix them. And now i need to fix ruby and remove those subpkg's 2013-09-13 11:18:52 does subpkgs has its own .gemspec ? 2013-09-13 11:19:41 and to make redmine possible (which deps on older ruby pkg's (rails) I am creating dedicated redmine ruby packages to install lower versions to make redmine happy. 2013-09-13 11:20:48 vkrishn: http://git.alpinelinux.org/cgit/aports/tree/main/ruby/APKBUILD#n119 2013-09-13 11:21:26 to be honest, I dont know a lot of ruby's internals and its package manager (gems) 2013-09-13 11:23:35 you can still use apk and gem together. ruby allows mutliple version of a pacakges. 2013-09-13 11:24:59 hmm.. 2013-09-13 11:35:19 I also dont understand why json has been removed from ruby. 2013-09-13 11:36:08 maybe it conflicts with ruby-json 2013-09-13 12:51:36 by hassle on end-user pov, you mean for packages that have ext/ dir that needs compiling ? 2013-09-13 13:01:14 vkrishn: I mean in general. users are expecting things to work with *just* the package manager. 2013-09-13 13:01:56 but the audiance for alpine is probably a little different. 2013-09-13 13:08:27 if I install ruby, does that install gem app ? 2013-09-13 13:08:46 no 2013-09-13 13:08:58 but i think ruby needs it 2013-09-13 13:09:38 atleast last time i tried something, it was spitting errors about gem missing. 2013-09-13 13:10:52 once I install gem app, does that install all packages needed to build a gem ? 2013-09-13 13:11:03 no 2013-09-13 13:11:14 you need ruby-dev 2013-09-13 13:11:19 that will install the deps 2013-09-13 13:11:30 but you will probably need alpine-sdk also 2013-09-13 13:14:00 hmm.... than I guess present gems packaging would be preferable 2013-09-13 13:15:07 I think i fixed ruby 2013-09-13 13:15:14 :-) 2013-09-13 13:15:26 spec file location changed 2013-09-13 15:22:51 that should be the last one.. 2013-09-14 08:45:30 more coffee, less commits 2013-09-14 08:46:14 barthalion: how would you build mono3? 2013-09-14 08:46:26 mono3-llvm? 2013-09-14 08:46:34 not sure if I would at all 2013-09-14 08:46:58 but we have mono 3 in Arch, so it's not impossible 2013-09-14 08:47:18 what is wrong with mono 3.2.1 to prefer -llvm over it? 2013-09-14 08:48:40 It's a performance boost for mono 2013-09-14 08:49:14 sigh 2013-09-14 08:49:28 "But computationally intensive applications will definitely benefit from this. Financial and scientific users will surely appreciate this performance boost." http://www.mono-project.com/Mono_LLVM 2013-09-14 08:49:29 shafire: does it have a stable release? 2013-09-14 08:49:39 only for mono2 :( 2013-09-14 08:50:23 https://github.com/mono/llvm 2013-09-14 08:50:29 I'll take a look at it after I fix build errors 2013-09-14 08:51:42 shafire: well, we don't have mono at all in aports, so I see no reason why not use mono 2 :p 2013-09-14 08:56:52 huh 2013-09-14 08:57:26 ah 2013-09-14 08:57:40 barthalion: mono2 is not latest stable version anymore 2013-09-14 08:58:58 oops, typo 2013-09-14 08:59:17 mh barthalion, how about taking mono3 and add the llvm diff as patch? 2013-09-14 08:59:21 http://www.mono-project.com/Mono:Runtime:Documentation:LLVM#The_LLVM_Mono_Branch 2013-09-14 08:59:30 shafire: we can upload a git snapshot to our servers and use it in APKBUILD 2013-09-14 08:59:38 if git sources work of course 2013-09-14 09:01:55 puf, forgot about that one 2013-09-14 09:02:46 barthalion: I think, the patch idea is not bad :D 2013-09-14 09:06:23 shafire: it may be not as easy as it seems to be, but I'm usually pessimistic :p 2013-09-14 09:10:14 algitbot: retry 2013-09-14 09:10:20 algitbot: build master 2013-09-14 09:15:13 Am I right that there no kvm packages? 2013-09-14 09:17:19 nailed it 2013-09-14 09:17:23 algitbot: in your face 2013-09-14 09:18:09 shafire: kvm is a kernel module 2013-09-14 09:18:55 shafire: after loading it, you add -enable-kvm to qemu arguments 2013-09-14 09:19:21 ah ok 2013-09-14 09:58:55 barthalion: I am happy to help 2013-09-14 10:03:17 I'll take a look in the evening if I won't be drunk too much, or tommorow 2013-09-14 10:03:59 ok 2013-09-14 10:14:02 there is no exfat package at moment, or? 2013-09-14 10:15:17 you're right 2013-09-14 10:17:29 http://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB#Manual <- I think, exFat is a better choice 2013-09-14 10:21:57 Can http://wiki.alpinelinux.org/wiki/Alpine_local_backup save my data directly on the bootable usb stick? 2013-09-14 10:22:34 possibly 2013-09-14 10:22:49 I usually perform normal installation on usb stick 2013-09-14 10:23:38 I want to run it from tmpfs 2013-09-14 10:24:00 you should ask ncopa 2013-09-14 10:24:41 yeah, we talked about it. But I can not find the log anymore, if I need 2 partitions or I can put it all in one :) 2013-09-14 10:24:46 ncopa: read here on monday 2013-09-14 10:25:42 I will create a feature request for exfat-utils 2013-09-14 10:25:47 fuse-exfat is not needed, or? 2013-09-14 10:26:19 I doubt you will boot anything using fuse :p 2013-09-14 10:26:27 exfat-utils should do 2013-09-14 10:36:01 I created two feature requests 2013-09-14 19:42:44 yeah i don't know how to fix libreswan 2013-09-15 09:27:22 hi 2013-09-15 13:41:06 hi 2013-09-16 08:03:01 hi 2013-09-16 08:03:47 ncopa: Have you seen #2253 already? 2013-09-16 08:03:47 shafire: hi 2013-09-16 08:04:42 why should we switch from vfat to exfat? 2013-09-16 08:05:45 there is no intree kernel support for it 2013-09-16 08:05:54 exfat is optimized for usb flash devices 2013-09-16 08:06:13 only fuse would currently be an option 2013-09-16 08:06:39 could you explain what "optimized for usb flash" means? 2013-09-16 08:06:50 https://github.com/dorimanx/exfat-nofuse 2013-09-16 08:06:52 or if you have any links, explaining the technical details 2013-09-16 08:07:41 why is the exfat-nofuse in mainline kernel? 2013-09-16 08:08:05 you mean isn't? 2013-09-16 08:08:07 files from v2.4.11-94-g31c25ec uploaded 2013-09-16 08:08:13 i think it is not... 2013-09-16 08:08:16 but i might be wrong 2013-09-16 08:08:56 "Big thanks to lqs for completing the driver! Big thanks to benpicco for fixing 3.11.y compatibility!" 2013-09-16 08:09:09 sounds like its not in mainline 2013-09-16 08:09:17 it isnt 2013-09-16 08:09:38 samsung recently released it under gpl 2013-09-16 08:09:39 they normally have good reasons to reject code in mainline kernel 2013-09-16 08:09:52 but there are lots of discussions about it. 2013-09-16 08:10:19 are there any technical reasons that it got rejected? 2013-09-16 08:10:49 its not yet send to linus afaik 2013-09-16 08:10:53 ok 2013-09-16 08:11:32 but we could add the utils for the fuse version 2013-09-16 08:11:39 yes 2013-09-16 08:11:57 so people could atleast access their sdcards 2013-09-16 08:12:32 if its going in the mainline kernel direction then we could even consider add it as a 3rd party module 2013-09-16 08:13:05 the fuse version could likely go in as is, now 2013-09-16 08:13:09 what i read, that could be a long way from happening. 2013-09-16 08:13:21 clandmeter: where you reading it? 2013-09-16 08:13:22 http://www.tech-recipes.com/rx/2801/exfat_versus_fat32_versus_ntfs/ 2013-09-16 08:13:34 samsung vs mircosoft vs linux. thats not a good combination... 2013-09-16 08:16:52 looks like main reason to do exfat is 32GB size limit and max filesize of 4G in fat32 2013-09-16 08:17:24 and faster: http://blog.testfreaks.com/information/usb-flash-drive-comparison-part-2-fat32-vs-ntfs-vs-exfat/ 2013-09-16 08:18:18 exfat filesize limit is not 32gb? 2013-09-16 08:18:42 It's 16 exabytes 2013-09-16 08:19:28 i'm talking about fat32 limits 2013-09-16 08:20:01 sorry 2013-09-16 08:21:43 i'm ok with adding the fuse driver 2013-09-16 08:22:11 i might be ok with adding it as a 3rd party driver (not sure about that yet though) 2013-09-16 08:22:36 but i'm not yet ok with using it as the default fs for boot usb 2013-09-16 08:23:02 Ok 2013-09-16 08:23:37 But I could use exfat as fs for my boot usb, or? 2013-09-16 08:23:37 let upstream decide what is best. no need to worry about it here. 2013-09-16 08:23:58 clandmeter: exactly... 2013-09-16 08:24:09 shafire: not yet 2013-09-16 08:24:42 i think to get that working, we need to first test the fuse dirver, then maybe with an updated initarmfs image it might be possible 2013-09-16 08:34:11 ncopa: does algitbot have zeromq interface? 2013-09-16 08:34:30 clandmeter: yes 2013-09-16 08:34:33 i can push notifications to it and report it here? 2013-09-16 08:34:45 yes 2013-09-16 08:35:10 can you point me in the right direction? 2013-09-16 08:39:21 how you did that? =) 2013-09-16 08:40:52 algitbot: give me a beer 2013-09-16 08:47:30 oa... 2013-09-16 08:47:41 that linux-grsec was not supposed to got here... 2013-09-16 08:48:30 now we see, if you broke it :) 2013-09-16 08:48:47 big patch 2013-09-16 08:48:52 it will break... 2013-09-16 08:48:56 on 3rdparty mods 2013-09-16 08:50:27 i think i'll reach push the 3rd party mods before kernel is built 2013-09-16 09:42:03 ncopa: I had a question too you 2013-09-16 09:42:08 too=for 2013-09-16 09:43:50 http://irclogger.com/.alpine-devel/2013-09-14#1379154286 2013-09-16 09:44:33 For the bootable usb stick, do I need two partitions (1 partition for the setup-bootable, and second for the lbu commits) or can I put both on one partition? 2013-09-16 09:45:42 shafire: you can have both on same partition 2013-09-16 09:46:29 ok, perfect. 2013-09-16 10:07:07 ncopa: I hope I got them all in #2256 im not sure about the last one though, but it doesnt harm. 2013-09-16 10:55:46 that's nice 2013-09-16 10:56:05 If someone want nimrod, there c-sources to create a bootstrap nimrod compiler 2013-09-16 10:56:11 there're 2013-09-16 10:56:40 www.nimrod-code.org 2013-09-16 11:05:57 who wanted the rust compiler? 2013-09-16 11:13:06 i am playing with the thought of convering the shell apkbuilds to lua 2013-09-16 11:13:56 why lua? :) 2013-09-16 11:15:02 Someone said, that in Lua, there is a table with variables, which you can access and manipulate the variables directly. 2013-09-16 11:18:12 because Lua is small and fast 2013-09-16 11:21:12 example aport in lua: http://sprunge.us/SZTK 2013-09-16 11:21:41 i think the biggest win will be how subpackages and split functions are defined 2013-09-16 11:21:57 drawback is that user might need learn both Lua and shell... 2013-09-16 11:22:19 but Lua will only/mostly be used for defining the things 2013-09-16 11:23:56 and dependency info 2013-09-16 11:28:08 zmq3 and ruby dont like each other... 2013-09-16 11:29:31 i guess that native redmine plugin will need to wait. 2013-09-16 12:26:46 look like shafire needs to buy some beer 2013-09-16 13:25:36 ncopa: why not split the maintainer string into name and email? 2013-09-16 13:26:35 thanks ncopa :) 2013-09-16 13:26:41 clandmeter: how can I do that? :)) 2013-09-16 13:27:58 shafire: just remember it. Maybe you will get the chance eventually. 2013-09-16 13:28:33 yeah 2013-09-16 13:29:02 how about the variable license to be a enum? 2013-09-16 13:29:04 an* 2013-09-16 13:29:14 and arch too 2013-09-16 13:32:48 oha 2013-09-16 13:33:00 enums in lua are not just numbers... 2013-09-16 13:33:35 shafire: its only at the idea stage still 2013-09-16 13:33:45 i dont even know if we want go that road 2013-09-16 13:35:30 I like the enums in Lua... 2013-09-16 13:35:47 clandmeter: what "zmq3 and ruby" issue ? 2013-09-16 13:36:27 ruby-zmq doesnt not support zmq3 2013-09-16 13:38:33 https://github.com/chuckremes/ffi-rzmq 2013-09-16 13:45:00 http://zeromq.org/bindings:ruby-ffi 2013-09-16 13:45:58 ncopa: how safe is setup-bootable? 2013-09-16 13:48:32 i think its about 42 :) 2013-09-16 13:48:50 vkrishn: thx ill looks into it 2013-09-16 13:49:24 what does that mean? 2013-09-16 13:49:33 42? 2013-09-16 13:49:55 what metric do you use to measure "safe"? :) 2013-09-16 13:50:20 :) 2013-09-16 13:51:02 ok if i break my remote box, ill start to inject harmfull code into grsec kernel ;-) 2013-09-16 13:54:17 42 is good 2013-09-16 13:54:27 watch out for syslinux-5 2013-09-16 13:54:36 its -u 2013-09-16 13:54:45 when? 2013-09-16 13:54:51 but this script doesnt really output anything usefull? 2013-09-16 13:54:57 algitbot: apk info syslinux 2013-09-16 13:55:01 -v 2013-09-16 13:55:38 clandmeter: if you setup-bootable -u from same stable series it should be safe 2013-09-16 13:55:54 why the -u? 2013-09-16 13:56:53 What was the impediment to not update to syslinux 6? 2013-09-16 13:57:16 ncopa: i first updated the box to 2.6 2013-09-16 13:57:37 it worked, so afterwards i used setup-bootable to update the kernel 2013-09-16 13:57:44 now i just need to reboot and pray 2013-09-16 13:57:55 ACTION prays with clandmeter :) 2013-09-16 13:57:56 clandmeter: be sure to use the "keep syslinux device" option 2013-09-16 13:58:06 clandmeter: without that flag, I've burned a couple remote boxes 2013-09-16 13:58:17 jbilyk_: doesnt -u do that? 2013-09-16 13:58:50 clandmeter: it's supposed to, but look at the description of -k 2013-09-16 13:58:54 eeks, it did update it... 2013-09-16 13:58:58 see :) 2013-09-16 13:59:07 clandmeter: without that flag, I've burned a couple remote boxes 2013-09-16 13:59:15 ^^ is from experience :D 2013-09-16 13:59:29 "hmmm...I'd better start packing for the 1000km drive" :D 2013-09-16 13:59:35 couldnt you have said that 2 min earlier? :) 2013-09-16 13:59:46 did you reboot? 2013-09-16 13:59:53 i was testing new AP setup in the office, and wasn't at desk lol 2013-09-16 13:59:59 sounds like i should pack to hide for a while... 2013-09-16 14:00:16 how hard is it to get a kvm? 2013-09-16 14:00:22 ncopa: I think -k should be default behaviour 2013-09-16 14:00:38 shafire: i thinkg its around 42 hard... 2013-09-16 14:00:46 -u Upgrade mode. Keep existing syslinux.cfg and don't run syslinux. 2013-09-16 14:01:16 i'm interested why/what breaks without the -k 2013-09-16 14:01:23 http://git.alpinelinux.org/cgit/alpine-conf/commit/?id=192c731713534451c4d5e76c0176a89a34ba6416 2013-09-16 14:01:32 i think it will replace the alpine_dev with uuid 2013-09-16 14:01:32 ncopa: apk cache has broken for me several times 2013-09-16 14:01:38 uuid should be safe... 2013-09-16 14:01:44 ah.. apk cache... 2013-09-16 14:01:45 -u Upgrade mode. Keep existing syslinux.cfg (but secretly modify if) and don't run syslinux. 2013-09-16 14:01:47 hm... 2013-09-16 14:02:04 yes i think thats the only change 2013-09-16 14:02:10 ncopa: broken apk cache has prevented several boxes from booting (or stopped firewall...or...etc) 2013-09-16 14:02:11 I hope its gona work 2013-09-16 14:02:23 hm.... 2013-09-16 14:02:27 ncopa: I've just learned to use -k :) 2013-09-16 14:02:39 jbilyk: care to file a bug for it? 2013-09-16 14:02:45 i think its worth use -k as the default... 2013-09-16 14:02:49 sure 2013-09-16 14:03:01 jbilyk_: why did the box fail? 2013-09-16 14:03:06 because of the uuid? 2013-09-16 14:03:13 i think apk cache gets wrong 2013-09-16 14:03:34 i think he likely had alpine_dev=sda1 2013-09-16 14:03:42 alpine_dev=usb or alpine_dev=usbdisk 2013-09-16 14:03:43 and pointed apk cache to /media/sda1/cache 2013-09-16 14:03:52 something like that, or sda1..can't remember 2013-09-16 14:04:04 and no /media/sda1 in fstab 2013-09-16 14:04:12 setup-bootable (w/o -k) changed it to 2013-09-16 14:04:18 and /media/ 2013-09-16 14:04:39 causing the /etc/apk/cache pointing to a non-existing location... 2013-09-16 14:05:18 seem my alix still has the correct cache symlink 2013-09-16 14:05:53 ncopa: #2258 created 2013-09-16 14:06:04 or it boots up without needing the apk cache 2013-09-16 14:06:21 jbilyk: thanks! 2013-09-16 14:06:39 clandmeter: I used to boot several boxes with CF card, but now that I use USB, UUID change breaks less things 2013-09-16 14:06:41 i think i'll invert it and change it to -U or something 2013-09-16 14:06:42 oh you mean when i reboot now it will set my media/sda1 to media/UUID=EC94-A0D2 ? 2013-09-16 14:07:10 err media/EC94-A0D2 2013-09-16 14:07:11 clandmeter: it could be. but it might it it just works too... 2013-09-16 14:08:37 ok ill reboot. the box is kind of local. 2013-09-16 14:09:04 just need to find a serial cable if it doesnt work 2013-09-16 14:09:28 "kind of local".. "i just just have to leave my chair and walk 5m" :) 2013-09-16 14:09:47 :) 2013-09-16 14:10:15 its not the walking, its to find a calbe and a serial port :) 2013-09-16 14:10:45 i think it will work 2013-09-16 14:13:10 It works 2013-09-16 14:13:41 I took a quick look in the future 2013-09-16 14:14:03 into* 2013-09-16 14:15:19 ncopa: why isnt there openssh inside alpine mini? 2013-09-16 14:15:51 atleast if i reboot and it cannot access my cache, it will still run ssh... 2013-09-16 14:15:59 hm 2013-09-16 14:16:14 good point 2013-09-16 14:17:32 im going to revert my syslinux file. just to make sure 2013-09-16 14:22:35 it worked :) 2013-09-16 14:23:02 so i think setup-bootables is safer than 42 then :) 2013-09-16 14:23:53 clandmeter: the reboot worked? 2013-09-16 14:24:09 yes 2013-09-16 14:24:13 are you running from tmpfs? 2013-09-16 14:25:07 yes 2013-09-16 14:25:33 :) 2013-09-16 14:27:06 almost all my boxes run from ram 2013-09-16 14:27:27 maybe, we should swap fuse and exfat in fuse-exfat. all other fuse packages are named so: http://alpinelinux.org/packages 2013-09-16 14:28:29 http://alpinelinux.org/packages?title=fuse 2013-09-16 14:31:46 ncopa: do you have an idea why my modules init fails to start? it doenst list any error except * ERROR: modules failed to start 2013-09-16 14:36:44 got it 2013-09-16 20:34:20 and let's watch how it fails in a couple of minutes 2013-09-16 20:40:18 awww 2013-09-16 20:40:21 no way! 2013-09-16 20:43:16 I'm really disappointed 2013-09-17 08:10:35 hi 2013-09-17 08:12:44 svn if broke :-/ 2013-09-17 08:12:47 is* 2013-09-17 08:12:57 svn http://.... 2013-09-17 08:25:19 hm, maybe it's my fault 2013-09-17 08:25:31 it may need rebuild against new serf 2013-09-17 08:35:17 well, but api nor abi didn't change 2013-09-17 09:18:42 no 2013-09-17 09:18:48 https://dpaste.de/4Pa7Z/ 2013-09-17 09:18:51 it tries ipv6 2013-09-17 09:18:57 but does not fallback to ipv4 2013-09-17 09:19:25 I've similiar issues with busybox wget sometimes 2013-09-17 09:27:55 echo 1 > /proc/sys/net/ipv6/conf/all/disable_ipv6 2013-09-17 09:31:28 anybody know how to import certs into android? 2013-09-17 09:32:09 seems as our startcom cert is not recognised in android 2013-09-17 09:32:31 clandmeter: startcom above android 4 2013-09-17 09:33:45 I have a nexus 4 2013-09-17 09:38:38 for some reason *cough*prism*cough* i no longer trust any CA's except my own... 2013-09-17 12:32:47 ncopa: can i output abuild to logfile? 2013-09-17 12:37:30 &> doesn't do the job? 2013-09-17 12:38:33 well, &> is probably bashism 2013-09-17 12:38:54 2>&1>logfile or something like that is POSIX 2013-09-17 12:39:07 hrm, no 2013-09-17 12:39:15 > logfile 2>&1 2013-09-17 13:00:09 barthalion: i was more looking for a abuild.conf solution 2013-09-17 13:00:15 i guess there is none 2013-09-17 13:00:30 or maybe and abuild option to sprunge it :) 2013-09-17 13:23:29 good afternoon 2013-09-17 13:23:38 clandmeter: yes it can 2013-09-17 13:23:53 abuild -m > file.txt 2>&1 2013-09-17 13:23:56 can please somebody try to pull from git://git.alpinelinux.org/fab/aports 2013-09-17 13:24:02 i think -m disables colors 2013-09-17 13:24:38 there are some new aports and some leftovers (poppler!) 2013-09-17 13:25:01 thanks 2013-09-17 13:27:11 my poppler upgrade is still there 2013-09-17 13:28:00 will poppler break things? 2013-09-17 13:29:04 barthalion suggested first to consult with you about those stuff 2013-09-17 13:31:55 ncopa: the changelog doesn't mention any api changes 2013-09-17 13:33:51 poppler tend to break abi every release. i think there was a drop of gtk2 support too 2013-09-17 13:34:10 so we might need upgrade evince 2013-09-17 13:34:17 dunno what happens with inkscape 2013-09-17 13:34:57 have you test built it? 2013-09-17 13:35:10 we have a checkapk script to verify ABI breakages 2013-09-17 13:35:21 if ABI breaks all apps linking to it will need to be rebuilt 2013-09-17 13:36:57 fabian_a: if you want i can pull your branch and cherry pick the stuff except poppler (which i am pretty sure will break things) 2013-09-17 13:37:42 well or i revert the poppler commits 2013-09-17 13:40:20 ncopa: is there a oneliner to create an signed apk index? 2013-09-17 13:58:57 apk index -o APKINDEX.tar.gz *.apk && abuild-sign APKINDEX.tar.gz 2013-09-17 14:11:21 ah ok 2013-09-17 14:11:29 wiki had some extra steps 2013-09-17 14:24:11 fabian_a: i am cleaning up / rebasing your commits 2013-09-17 14:24:25 removing the poppler things 2013-09-17 14:24:39 there are some testing/unpaper changes 2013-09-17 14:27:00 sorry, unpaper is messy 2013-09-17 14:27:00 i think I'd like to clean up those 2013-09-17 14:27:11 its supposed to be 0.3 right? 2013-09-17 14:27:17 i would like to have it clean 2013-09-17 14:27:30 git rebase -i is your friend ;) 2013-09-17 14:27:33 (read about it= 2013-09-17 14:27:48 basically 2013-09-17 14:28:06 yes, 0.3. there is a new fork of unpaper available but not building 2013-09-17 14:28:09 it looks like removing all commits that touches unpaper 2013-09-17 14:28:29 will reset it to the state where 0.3 is the current 2013-09-17 14:28:32 basically no changes 2013-09-17 14:29:17 after merging all the changes into one single commit i get this: https://dpaste.de/moTr4/ 2013-09-17 14:29:28 which is basically a no-op 2013-09-17 14:29:46 so i think i just remove all the unpaper commits 2013-09-17 14:30:40 would tat be ok? 2013-09-17 14:30:43 that* 2013-09-17 14:32:50 it's ok to remove everything after 1367bc1c35 2013-09-17 14:33:53 yeah 2013-09-17 14:34:33 90be6665602af99730defee2e70a502c83fccc45 main/py-imaging: minor layout modificat 2013-09-17 14:34:40 that change does not bump pkgrel 2013-09-17 14:34:58 but there should not really be any changes in the final .apk right? 2013-09-17 14:35:24 well, i didn't bump the pkgrel when there is no need for a rebuild 2013-09-17 14:35:31 good 2013-09-17 14:35:37 yes its correct 2013-09-17 14:39:02 pick 5dc73d6 testing/py-bottle: duplicate 2013-09-17 14:39:02 pick 90e171a testing/py-bottle: duplicate 2013-09-17 14:39:09 second commit is a fixup right? 2013-09-17 14:39:11 looks so 2013-09-17 14:39:29 i'll clean that up... 2013-09-17 14:39:42 i push bottle to my repo but it was included in the meantime 2013-09-17 14:39:50 thanks 2013-09-17 14:39:54 i'll fix it 2013-09-17 14:41:19 4933cccce15eb9ff286c43a049d1bf999f96428b 2013-09-17 14:41:24 testing/py-htmlcolor: new aport 2013-09-17 14:41:31 +pkgname=py-pygtail 2013-09-17 14:41:32 +_pkgname=pygtail 2013-09-17 14:41:52 should it be py-htmlcolor or py-pygtail? 2013-09-17 14:42:12 the dirname does not correspond with pkgname 2013-09-17 14:43:32 oops, this went to the wrong APKBUILD 2013-09-17 14:43:48 i will fix that 2013-09-17 14:43:53 how should it be? 2013-09-17 14:44:03 umm i have a rebase already... 2013-09-17 14:44:31 its probably easier for me to fix this here than do a new pull and fix all the above once again... 2013-09-17 14:45:23 ok, pkgname=py-htmlcolor 2013-09-17 14:45:30 _pkgname=htmlcolor 2013-09-17 14:45:45 pkgver=0.1 2013-09-17 14:45:56 license="BSD" 2013-09-17 14:46:18 url="http://github.com/ext/pyhtmlcolor" 2013-09-17 14:46:59 pkgdesc="A Python parser for HTML and CSS color values" 2013-09-17 14:49:02 super 2013-09-17 14:49:04 thanks! 2013-09-17 14:50:52 now buildtesting can start... 2013-09-17 14:51:16 this was a very inefficient way to create a pkg ;-) 2013-09-17 15:00:32 aw gotta go 2013-09-17 15:00:34 but 2013-09-17 15:00:44 ruby-typhoeus 2013-09-17 15:00:46 is disabled 2013-09-17 15:00:52 can i remove it alltogether? 2013-09-17 15:02:14 i just did 2013-09-17 15:03:16 i hope it does not break anything. i believe it won't 2013-09-17 15:03:18 thanks! 2013-09-17 15:03:20 see u 2013-09-17 18:53:53 hi 2013-09-17 18:53:59 fabian_a: will it run on cubieboard too? 2013-09-17 20:18:49 shafire: sorry, i didn't get the point 2013-09-18 06:55:05 nctest: hello 2013-09-18 11:31:37 hi 2013-09-18 11:32:41 today it was very quite ;) 2013-09-18 11:32:59 busy here... 2013-09-18 11:36:29 fabled: debootstrap does not not work 2013-09-18 11:36:32 E: Invalid Release file, no entry for main/binary-uclibc-linux-amd64/Packages 2013-09-18 11:36:57 what i find interesting is why it tries to do *-uclibc-* in there 2013-09-18 11:37:16 my guess is that --host/--build option to deboostrap caused this 2013-09-18 11:39:02 hum... no! 2013-09-18 11:39:12 there are no ./configure to debootstrap 2013-09-18 11:45:28 sigh... debootstrap is broke again.. 2013-09-18 11:47:05 ah its dpgs 2013-09-18 11:47:18 ncopa-desktop:~# dpkg --print-architecture 2013-09-18 11:47:18 uclibc-linux-amd64 2013-09-18 11:49:26 yup, if i remove --host/--build from dpkg it will: 2013-09-18 11:49:32 $ dpkg --print-architecture 2013-09-18 11:49:32 i386 2013-09-18 11:52:08 Hi all! 2013-09-18 11:52:14 ncopa: thats debian right? for what do you use it? 2013-09-18 11:52:22 Could I have some assistance with someone (preferably as many as possible) to look at http://wiki.alpinelinux.org/wiki/Ensuring_that_/var_is_correctly_mounted_on_HDD to see if it looks ok, or if this could be done in a better/simpler way? 2013-09-18 11:52:52 clandmeter: so i can easily get a glibc install in an lxc 2013-09-18 11:52:57 to test/compare etc 2013-09-18 11:59:51 mhavela: there is a lot of information in there 2013-09-18 12:01:55 mhavela: it doesnt look structured, so its dificult to read. 2013-09-18 12:06:45 clandmeter, Thanks for your input! I try to see what I can do about it! 2013-09-18 12:06:49 Thanks for having a look! 2013-09-18 12:07:18 mhavela: you target end-users right? 2013-09-18 12:07:41 Yes 2013-09-18 12:08:02 um.. 2013-09-18 12:08:06 maybe you should try to prevent using awk in your cmd's 2013-09-18 12:08:17 so people know what is happening 2013-09-18 12:08:21 Good thought 2013-09-18 12:08:49 i dont think there are any users that are "affected" unless they have chosen set up their box that way in the first place... 2013-09-18 12:09:34 ncopa, Changed the subheading. 2013-09-18 12:09:40 thats another point :) 2013-09-18 12:09:52 I got things wrong when saying that there was a early script doing things the other way around 2013-09-18 12:09:53 and im not sure i would fix it like that 2013-09-18 12:10:34 its also very specific, because at the end you talk about raid devices. 2013-09-18 12:10:56 but thats kind of what ncopa said 2013-09-18 12:12:17 maybe you should put the conversion of filesystem and creating raid devices in its own page (if there isnt any already) 2013-09-18 14:53:21 ncopa: checkapk is broken? 2013-09-18 15:02:41 dunno 2013-09-18 15:06:28 i thought it was as simple as abuild -r && checkapk? 2013-09-19 06:21:20 clandmeter: its supposed to be just abuild -r && checkapk 2013-09-19 06:21:35 you need atleast one http repo 2013-09-19 07:37:01 ncopa: i have repo's added, but it msg it cannot find the apk 2013-09-19 07:38:28 checkapk: can't find v4l-utils-1.0.0-r0.apk 2013-09-19 08:06:16 apk update? 2013-09-19 08:06:44 run sh -x /usr/bin/checkapk 2013-09-19 08:06:47 to find where it fails 2013-09-19 08:08:39 ah it doesnt add the arch 2013-09-19 08:09:02 so its a bug in checkapk? 2013-09-19 08:09:37 im checking 2013-09-19 08:09:53 + [ -f /home/clandmeter/packages//main//v4l-utils-1.0.0-r0.apk ] 2013-09-19 08:10:18 yes 2013-09-19 08:10:22 it fails to detect the arch 2013-09-19 08:10:30 smells like checkapk bug 2013-09-19 08:10:58 yup 2013-09-19 08:11:25 due to the removal of CARCH from abuild.conf 2013-09-19 08:16:32 clandmeter: could you please test this: http://sprunge.us/bEHG 2013-09-19 08:21:41 ncopa: seems to work 2013-09-19 08:22:06 still needs cleanup 2013-09-19 08:22:19 i think the logic of setting CARCH should all be moved ot functions.sh 2013-09-19 08:22:27 then we get it for free for all scripts 2013-09-19 08:28:46 ncopa: http://sprunge.us/Zjja 2013-09-19 08:28:52 does that say anything to you? 2013-09-19 08:36:19 the sigsev? 2013-09-19 08:36:26 looks liek a bug in the usage printing 2013-09-19 10:43:18 Hey all, I have a serious doubt that quagga zebra and the like routing daemons, should be started after firewall during boot time .. 2013-09-19 14:13:46 i just went ahead and put it in main because i've been using it for a while 2013-09-19 14:24:44 kaniini: you never use newapkbuild ? 2013-09-19 14:27:08 clandmeter: i do 2013-09-19 14:29:15 kaniini: i though with the latest changes to aports, we need --build and --host. 2013-09-19 14:31:05 no 2013-09-19 14:31:29 only for packages requiring specific help cross-compiling 2013-09-19 14:31:42 fish's build system will pick it up from $CC 2013-09-19 14:33:48 ok, i wasnt sure so i just asked :) 2013-09-19 14:36:02 anyway, fish is cool 2013-09-19 14:36:07 it has rc(1)-like syntax 2013-09-19 14:40:11 yes i looked it up 2013-09-19 14:40:17 i should give it a try 2013-09-19 14:41:01 if it wasn't C++ bloatware (at 780kb), i would even propose making it the default system shell 2013-09-19 14:41:03 lol 2013-09-19 14:41:09 (for logins) 2013-09-19 14:44:02 kaniini: it screams for a fork in c ;-) 2013-09-19 14:44:25 nah 2013-09-20 07:32:52 I don't seem to be able to install 'subversion' on a box that uses edge and 2.6 as its repository 2013-09-20 07:33:08 (This is a LXC containered box, but I dont think that matters) 2013-09-20 07:33:14 It saids 2013-09-20 07:33:29 ERROR: unsatisfialbe constraints: 2013-09-20 07:33:37 so:libserf-1.so (missing): 2013-09-20 07:34:02 required by: subversion-libs-1.8.3-r0[so:libserf-1.so] 2013-09-20 07:34:03 mhavela: ncopa said svn is broken 2013-09-20 07:34:04 ... 2013-09-20 07:34:12 Aah. 2013-09-20 07:34:14 i guess he didnt fix it yet 2013-09-20 07:34:17 ok. Thanks 2013-09-20 07:34:47 That means he knows about it already - which is good 2013-09-20 07:35:03 iirc it's brokem due to other bug 2013-09-20 07:38:14 barthalion: that errer he has seems like abi change? 2013-09-20 07:38:22 error 2013-09-20 07:43:18 yeah, we can try rebuild 2013-09-20 07:43:34 howerer I haven't touched 2.6 2013-09-20 07:43:41 however* 2013-09-20 07:46:51 i cant even build it local 2013-09-20 07:49:05 he, i had it installed, but when reinstalling it fails :) 2013-09-20 07:49:52 mhavela: try to install it when your mirror catch up with changes 2013-09-20 07:50:21 barthalion: works now 2013-09-20 07:50:33 ncopa: I told you we need a rebuild :p 2013-09-20 07:50:48 he's hiding 2013-09-20 07:51:00 surely he is! 2013-09-20 07:53:18 barthalion, Works just great now! 2013-09-20 07:53:19 Thanks! 2013-09-20 07:56:26 np :) 2013-09-20 07:56:28 anybody know how to stop awall/iptabled from reporting dropped packets to dmesg? 2013-09-20 07:56:50 I cant get my head around its documentation... 2013-09-20 08:00:31 clandmeter, you need to alter logging class named _default 2013-09-20 08:01:24 "Default logging settings can be set by defining a logging class named _default. Normally, default logging uses the log mode with packets limited to one per second." 2013-09-20 08:02:15 yes i read that part 2013-09-20 08:03:35 or each policy/filter rule may have the attribute "log": "false" to disable logging 2013-09-20 08:04:59 so toplevel i need to sepcify "log" and inside i need to define _default? 2013-09-20 08:05:50 rnalrd: do you have a sample so i can visualise it? 2013-09-20 08:06:15 http://pastebin.com/Pa9XrpHk 2013-09-20 08:11:24 rnalrd: netflow is defined somewhere else in your config? 2013-09-20 08:12:02 I can change that to _default i guess? 2013-09-20 08:12:42 yes, you can replace "netflow" with "_default" and all policy and rules will take that settings as default 2013-09-20 08:12:54 policy and filters* 2013-09-20 08:15:45 rnalrd: thx! 2013-09-20 08:15:57 yw 2013-09-20 08:31:19 Hey all, it seems to me that I have a small problem here in 2.6.4, ip -s route show cache shows me nothing .. :-( 2013-09-20 08:44:57 Has anybody reported that ACF is broken in 'edge'? 2013-09-20 08:45:14 rnalrd: i tried to disable logging completely like: "log":{"_default":"false"} 2013-09-20 08:45:18 I just brought up a new box and I get this when trying to browse to ACF: 2013-09-20 08:45:20 i guess thats not supported? 2013-09-20 08:45:21 mhavela: I guess no 2013-09-20 08:45:29 haserl CGI Error Lua shells are not enabled. 2013-09-20 08:45:58 clandmeter, that doesn't sound right 2013-09-20 08:46:24 lua doesnt like it 2013-09-20 08:46:38 "log": { "_default": { "log": "false" } } 2013-09-20 08:47:09 not sure this is going to work though 2013-09-20 08:47:49 you might want to add "log": "false" attribute for each "drop" or "reject" "action" instead 2013-09-20 08:48:14 Hey all, it seems to me that I have a small problem here in 2.6.4, ip -s route show cache shows me nothing .. :-( 2013-09-20 08:48:34 rnalrd: yes i think this doesnt work according iptables rules 2013-09-20 08:48:53 I think somethings wrong with haserl (on my working box I have haserl-0.9.30-r0 and on this broken box I have haserl-0.9.31-r0) 2013-09-20 08:49:01 Guess I should write a bugreport on this 2013-09-20 08:51:35 grharry: same about your issue, bugtracker is better place if no one answer here 2013-09-20 08:52:58 who invented json anyway? 2013-09-20 08:53:19 couldnt they made it noob friendly? 2013-09-20 08:56:16 and who would use fancy json parsing libraries then? 2013-09-20 09:19:36 barthalion: i want a json-interface-for-dummies :) 2013-09-20 09:21:12 and dont be soo stict on commas and bolean encapsulation 2013-09-20 12:55:50 clandmeter: ping 2013-09-20 12:56:08 jbilyk_: pong 2013-09-20 12:56:16 clandmeter: (finally) looking at www.a.o backups 2013-09-20 12:56:29 :) 2013-09-20 12:56:36 first issue I see is that /root/nightly-alpine-www.sql is *old* 2013-09-20 12:56:38 it's from may 2013-09-20 12:56:50 is that where nightly mysql backups are being dumped still? 2013-09-20 12:57:04 i didnt setup back-up 2013-09-20 12:57:09 ok 2013-09-20 12:57:14 so im not sure 2013-09-20 12:57:22 it uses the same db as bugs 2013-09-20 12:57:32 but i guess you do seperate db exports right? 2013-09-20 12:57:45 yeah, just saw the /etc/lbu/pre-package.d/ that exports the db 2013-09-20 12:57:48 doenst the script dump the b? 2013-09-20 12:57:55 yeah, it does 2013-09-20 12:57:58 but it must be failing 2013-09-20 12:57:59 ok 2013-09-20 12:58:07 but if it faisl it wont dump 2013-09-20 12:58:30 i just ran it manually and it works 2013-09-20 12:58:32 weird 2013-09-20 12:59:17 i think lbu had isues 2013-09-20 12:59:32 yeah, but I'm stepping through lbu's process to find why 2013-09-20 12:59:41 gzip: short write 2013-09-20 12:59:42 tar: write error: Broken pipe 2013-09-20 12:59:52 normally ^^ is diskspace issues, but you've got lots 2013-09-20 13:01:37 maybe tmp full? 2013-09-20 13:01:40 oh - your /tmp is mounted and is 16 2013-09-20 13:01:42 16M 2013-09-20 13:01:43 yeah 2013-09-20 13:02:42 clandmeter: is 16M /tmp because it's KVM ? 2013-09-20 13:02:54 i haven't used KVM much 2013-09-20 13:03:06 you dont have to 2013-09-20 13:03:23 KVM -> vserver -> guest 2013-09-20 13:04:11 oh yeah...forgot about how funky the setup is :) 2013-09-20 13:04:43 clandmeter: can we not mount /tmp separately, or give it 1G ? 2013-09-20 13:05:00 1G? 2013-09-20 13:05:04 are you serious? 2013-09-20 13:05:13 well, the backups are 170Mb 2013-09-20 13:05:23 why it uses so much tmp space? 2013-09-20 13:05:43 i'd give 1G just to have overhead in case we ever have big increase in drupal filesizes 2013-09-20 13:05:55 clandmeter: I'd be happy with 256MB for now :) 2013-09-20 13:06:03 clandmeter: at least that'd let backups work 2013-09-20 13:06:03 but it wil reserve it from system memory 2013-09-20 13:06:14 can we just not mount /tmp then? 2013-09-20 13:06:54 alternative is that I change backup script to manually make an apkovl instead of using lbu i guess 2013-09-20 13:07:00 then it wouldn't be writing to /tmp 2013-09-20 13:07:14 maybe that's simplest thing to do for now 2013-09-20 13:07:35 I can increase tmp, but im not sure how to mount it somwhere else 2013-09-20 13:07:56 clandmeter: you know what, don't worry about it - I'll just change the script to not use lbu 2013-09-20 13:08:04 ok 2013-09-20 13:08:17 it'll just be a one-off change, but at least we'll have backups :) 2013-09-20 13:08:33 do you put the db inside ovl? 2013-09-20 13:08:37 yeah 2013-09-20 13:08:56 you cannot seperate them? 2013-09-20 13:09:14 we could, but this way you have everything to recreate the www box in one spot 2013-09-20 13:09:21 clandmeter: drupal files are the big hog 2013-09-20 13:09:23 they're 100M 2013-09-20 13:09:35 im thinking about drupal 2013-09-20 13:09:36 so either way, we couldn't use lbu to backup with a 16M /tmp 2013-09-20 13:09:42 there aint much 2013-09-20 13:09:49 let me check 2013-09-20 13:09:54 alpine-www:~# du -hs /var/www/localhost/htdocs/ 2013-09-20 13:09:55 103.6M /var/www/localhost/htdocs/ 2013-09-20 13:10:05 clandmeter: ^ 2013-09-20 13:10:22 oh - you have multiple copies of drupal inside there :) 2013-09-20 13:10:30 yes 2013-09-20 13:10:38 and only sites is interesting 2013-09-20 13:10:53 /var/www/localhost/htdocs/drupal/sites 2013-09-20 13:11:17 clandmeter: that's still 34.5M 2013-09-20 13:11:26 not if you compress it i guess 2013-09-20 13:11:35 any tools to manage lxc containers ? 2013-09-20 13:12:01 clandmeter: I'll just tar.gz what we were doing in lbu 2013-09-20 13:12:40 sites and db is important 2013-09-20 13:12:44 the rest i dont care 2013-09-20 13:13:03 ok 2013-09-20 13:13:16 those are the only thing unique 2013-09-20 13:13:30 the rest is setup in less then an hour 2013-09-20 13:14:11 can't you merge 2 files for backup 2013-09-20 13:14:17 yup, that's what I'll do 2013-09-20 13:14:22 I'll use lbu to grab /etc 2013-09-20 13:14:26 and then add in db and www file 2013-09-20 13:14:28 files* 2013-09-20 13:14:29 one is db and sites and other is ovl 2013-09-20 13:14:48 then everybody can restore it 2013-09-20 13:14:52 yup 2013-09-20 13:16:09 even ppl from Canada 2013-09-20 13:16:13 :P 2013-09-20 13:16:19 :) 2013-09-20 13:18:18 clandmeter: ok, backups are working now 2013-09-20 13:18:42 clandmeter: /home/backups/*.apkovl.tar.gz are for /etc and www-files.*.tar.gz are www files and db 2013-09-20 13:18:50 /var/www/localhost/htdocs/drupal/sites to be precise 2013-09-20 13:19:01 done from /etc/periodic/daily/daily-backup 2013-09-20 13:21:40 ncopa: ^^ www.a.o backups are working again 2013-09-20 13:22:16 bummer. it seems gcc has armv6 related bugs. 2013-09-20 13:22:40 i think i'll retry with pure armv7 toolchain. this means rpi support will be delayed, as it's armv6. 2013-09-20 13:22:47 jbilyk_: super! 2013-09-20 13:23:49 clandmeter: I've got a good copy on the backup server now too, so we're all set 2013-09-20 13:24:06 ACTION scratches that off my todo list 2013-09-20 13:24:16 sorry it took a while :) 2013-09-20 13:24:21 np 2013-09-20 13:24:25 we pay you later 2013-09-20 13:25:26 clandmeter: http://www.lcbo.com/gifts/en/giftcards.shtml# works, especially the "bulk orders" :D 2013-09-20 13:26:03 :) 2013-09-20 13:26:16 they dont have amazon in .ca? 2013-09-20 13:26:38 yeah, but amazon.ca seems to be more third party vendors nowadays than actual amazon 2013-09-20 13:26:51 last few times I got an order from them it was a third-party 2013-09-20 13:30:47 you can choose prime 2013-09-20 13:30:53 those are mostly from amazon 2013-09-20 13:31:03 clam-av does not build 2013-09-20 13:31:07 cheaper shipping 2013-09-20 13:31:40 fabled: config files moved? 2013-09-20 13:31:46 sed: /home/buildozer/aports/main/clamav/pkg/clamav/etc/clamav/freshclam.conf: No such file or directory 2013-09-20 13:31:47 sed: /home/buildozer/aports/main/clamav/pkg/clamav/etc/clamav/clamd.conf: No such file or directory 2013-09-20 13:31:47 seems like that 2013-09-20 13:32:02 who broke it? 2013-09-20 13:32:05 could he fix it? 2013-09-20 13:32:09 ACTION hides 2013-09-20 13:32:56 i think it's the clamav upgrade to 0.98 from Bartłomiej Piotrowski that broke it 2013-09-20 13:36:28 config files are now .sample's 2013-09-20 13:38:09 should we keep them as .sample, or rename them to .conf ? 2013-09-20 13:38:13 i think keep .sample 2013-09-20 14:07:55 fabled: seems many packages not yet musl compat? 2013-09-20 14:08:05 yeah - well usually it's a bug in the package 2013-09-20 14:08:14 mostly it's just updating autotools config.sub 2013-09-20 14:08:24 to recognize the new toolchain triplet 2013-09-20 14:08:33 that openssh seems more then just autottools 2013-09-20 14:09:08 yes, there's some of those too 2013-09-20 14:09:12 usually it's bug in the application 2013-09-20 14:09:22 not including right headers, or just using wrong stuff 2013-09-20 14:09:22 i guess building the whole tree is cumbersome 2013-09-20 14:11:23 clandmeter, yes, it'll be work 2013-09-20 14:11:41 i'm now pushing what i've done earlier to do bootstrap arm environment 2013-09-20 14:11:59 sabotage has patches to many apps; and it's getting better all along. 2013-09-20 14:12:09 some issues also do really need to be fixed in the application 2013-09-20 14:12:13 how does upstream react? 2013-09-20 14:12:41 most of them actually know the exstence of musl? 2013-09-20 14:16:34 huh 2013-09-20 14:17:14 oh, seems openssh got new config.sub upstream 2013-09-20 14:17:50 didn't realize it got updated under me 2013-09-20 14:23:16 urgh 2013-09-20 14:24:26 im glad that does happen to my apk's 2013-09-20 19:12:38 p/buffer 14 2013-09-21 10:45:23 So if I wanted to make Alpine rockin' on Apple hardware, would that be a welcome thing? I'm just asking before I invest a lot of time in to it whether the community would care (i.e.: I should write it all out and document it and package shit) or not (i.e.: hack it until I like it) 2013-09-21 19:03:51 foxkit: i'd say go for it 2013-09-23 13:52:01 fabled: did you fix your gcc issues? 2013-09-23 13:52:29 have not tested all details yet 2013-09-24 16:39:06 i broke ada on edge - sorry about that 2013-09-24 16:39:11 i'm working on fixing it 2013-09-24 16:39:40 will push soon, and as i need to recover using older gcc with working ada, i hope noone pushes anything to edge until i say so. 2013-09-24 16:39:46 sorry for the inconvenience 2013-09-24 16:43:17 i am sure the US Navy (who is the only people who use ada lets be honest here) will really be worked up 2013-09-24 16:43:19 :P 2013-09-24 16:47:12 urgh. my fix didn't work. :/ need to hold on with pushing things. 2013-09-24 16:47:19 edge is free to push until then 2013-09-24 16:50:05 blah 2013-09-24 16:50:09 it never worked right 2013-09-24 16:50:27 ok, at least i know the right fix now 2013-09-24 17:15:31 we are also going to setup build-edge-musl 2013-09-24 17:15:42 building on arm is slow 2013-09-24 17:15:55 so to fix all the musl issues we try to build it on x86 first 2013-09-24 17:16:15 so far the arm work is blocked by gcc misbehaving 2013-09-24 17:16:18 for armv6 2013-09-24 17:16:28 armv7-a target works (but will not support rpi) 2013-09-24 17:18:26 do we really care about armv6? 2013-09-24 17:23:27 well 2013-09-24 17:23:43 many seem to care about raspberry pi 2013-09-24 17:23:47 i have one here too 2013-09-24 17:23:50 then again 2013-09-24 17:23:53 armv7 is a lot better 2013-09-24 17:24:03 so it might even make sense to build separately for both 2013-09-24 17:24:07 armv7 in thumb2 2013-09-24 17:24:11 armv6 in arm mode 2013-09-24 17:33:38 ok 2013-09-24 17:33:41 pushing stuff now 2013-09-24 17:33:52 edge frozen until gcc fixed 2013-09-24 17:33:59 the first attempts will fail 2013-09-24 17:38:08 uh. it'll build firefox first too. 2013-09-24 17:39:15 aaand, firefox fails with: 2013-09-24 17:39:17 configure: error: gstreamer and gstreamer-plugins-base development packages are needed to build gstreamer backend. Install them or disable gstreamer support with --disable-gstreamer 2013-09-24 17:39:53 well, we have gstreamer in repos 2013-09-24 17:40:16 both 1.0 and 0.10, iirc Fx wants 0.10 2013-09-24 17:41:07 if you disable it, some media backends won't be available 2013-09-24 17:41:42 no reason to do so, but at the same time… how many users have Alpine as main desktop OS? :p 2013-09-24 17:41:59 me! 2013-09-24 17:42:03 and ncopa 2013-09-24 17:42:16 oh, I knew only about ncopa :D 2013-09-24 17:42:18 barthalion: I've set it up in a few locations 2013-09-24 17:42:28 laptop, media player and a test desktop 2013-09-24 17:42:45 oh, and a streaming server so that local users could see the video as it was streaming 2013-09-24 17:43:06 fabled: let me see if I can fix it 2013-09-24 17:43:10 i'm on it 2013-09-24 17:43:46 let's see 2013-09-24 17:43:54 if that does not work, i'll revert the ugprade, i think 2013-09-24 17:50:47 that was gcc 2013-09-24 17:50:53 build-edge should do better 2013-09-24 17:51:22 will retry build64-edge soonish 2013-09-24 18:19:06 build64-edge: retry 2013-09-24 18:22:59 what's the command now adays to force builder to retry? 2013-09-24 18:25:01 lgitbot: build build-edge 2013-09-24 18:25:04 algitbot: build build-edge 2013-09-24 18:25:20 algitbot: build master 2013-09-24 18:33:40 oookay... should recover. feel free to push to edge again. 2013-09-24 19:04:07 fsck 2013-09-24 19:09:36 turns out it didn't fix all of the ada problems 2013-09-24 19:09:38 weird 2013-09-24 19:09:58 basic ada stuff works - but build gcc doesn't 2013-09-24 19:10:08 apparently it aborts on exceptions 2013-09-24 19:10:26 sounds like issue in uclibc/libgnats/libgcc interaction in unwinding stack. yey. 2013-09-24 19:22:33 will look at it tm 2013-09-24 21:01:57 please git pull git://git.alpinelinux.org/fab/aports 2013-09-24 21:02:26 there is nikola a modular, fast, simple, static website generator and its deps 2013-09-25 04:01:55 urgh 2013-09-25 04:02:00 libgcc is utterly broken in gcc now 2013-09-25 04:03:44 ah 2013-09-25 04:03:57 i know why - it's the musl patch - because gcc configure is utterly stupid 2013-09-25 04:04:26 :D 2013-09-25 04:04:58 sign 2013-09-25 04:05:57 Mem: 75487720k total, 1866604k used, 73621116k free CPUs: 48 @ 2266MHz 2013-09-25 04:06:05 just deployed alpine-xen on a bunch of nodes like this 2013-09-25 04:06:19 yeehaw 2013-09-25 04:07:06 cool. i have some 24 core Xeons boxes out there as well. 2013-09-25 04:18:54 mmm 2013-09-25 04:18:57 i'll just push 2013-09-25 06:58:15 ncopa, should we add "-l 8.0" MAKEFLAGS in build.alpinelinux.org's vservers ? 2013-09-25 06:59:01 whats -l 8.0 ? 2013-09-25 06:59:20 limit jobs based on cpu load 2013-09-25 06:59:47 so if two vservers build simultaneous, it wont spam max processes on all vservers 2013-09-25 07:00:31 ah 2013-09-25 07:00:37 sounds like what we want 2013-09-25 07:01:16 ok i think i found the bus-error issue 2013-09-25 07:01:21 gcc code is broke. again. 2013-09-25 07:05:51 or more like "-l ${JOBS}.0" 2013-09-25 07:05:53 or so 2013-09-25 07:27:10 ok 2013-09-25 07:27:16 got musl-x86 bootstrap done 2013-09-25 07:27:19 seems to work 2013-09-25 07:48:54 :wq 2013-09-25 08:54:04 ncopa: did you ever make usb devices available for vserver guests? 2013-09-25 08:55:04 no 2013-09-25 08:55:35 you mean like hotplugging? 2013-09-25 08:55:44 or a single usb disk 2013-09-25 08:55:51 single usb device 2013-09-25 08:56:00 disk? 2013-09-25 08:56:03 or something else? 2013-09-25 08:56:07 I would like to have a card reader 2013-09-25 08:56:24 inside the guest 2013-09-25 08:56:46 and it should be mounted as long as the vserver runs? 2013-09-25 08:56:51 or hotpluggable 2013-09-25 08:56:58 so you want mount if from the vserver guest 2013-09-25 08:57:35 what do you mean with hotplubable? 2013-09-25 08:57:37 g 2013-09-25 08:57:56 i mean if you want be able to unplug the card while its running 2013-09-25 08:57:57 i can just reinsert it and it would be available again? 2013-09-25 08:58:07 there is no need to remove it. 2013-09-25 08:58:10 thats the question 2013-09-25 08:58:11 ok 2013-09-25 08:58:15 if no need to remove it 2013-09-25 08:58:18 then its fairly simple 2013-09-25 08:58:29 mount sys dir in guest? 2013-09-25 08:58:51 oh card reader... 2013-09-25 08:58:59 i was thinking SD cards... 2013-09-25 08:59:05 :) 2013-09-25 08:59:08 so basically a mounted disk.. 2013-09-25 08:59:14 no sd card 2013-09-25 08:59:24 its a sattelite card reader 2013-09-25 08:59:46 maybe ask on OFTC #vserver 2013-09-25 08:59:53 i havent done it 2013-09-25 08:59:56 i believe it is possible 2013-09-25 09:00:41 maybe bind mount /sys/bus/usb or similar? 2013-09-25 09:00:53 how is it accessed? 2013-09-25 09:01:00 via /sys or via a /dev/ ? 2013-09-25 09:01:57 let me check 2013-09-25 09:04:21 /dev/ttyUSB0 2013-09-25 09:05:35 clandmeter: then i think it should be enough to just cp -a it to /vservers//dev/ 2013-09-25 09:07:59 ncopa, i want to backport the above cherry-pick for asterisk; asterisk seems to have nologin changes too; should i cherry-pick those apkbuild changes to 2.6-stable while at it? 2013-09-25 09:08:14 i think not 2013-09-25 09:08:32 we can leave the nologin change to v2.7 2013-09-25 09:08:35 ok 2013-09-25 09:18:58 ncopa, is the cat captcha working? i can't edit my wiki page. 2013-09-25 09:21:14 i wanted to update the bootrapping stuff 2013-09-25 09:21:42 there was some report that it was broke 2013-09-25 09:21:46 i dont know if it was fixed 2013-09-26 04:41:06 is it just me or is syntax highlighting in vim in edge not working anymore 2013-09-26 04:50:24 ok it was just me 2013-09-26 04:50:30 rebooted a couple of times and now it works.. lol 2013-09-26 06:15:35 ncopa: for developing any tools for build process at AL, what would be preferred scripting lang ? python, lua ? 2013-09-26 06:17:43 shell or lua 2013-09-26 06:18:45 yes, shell or lua 2013-09-26 06:20:30 http://wiki.alpinelinux.org/wiki/Setting_up_a_compile_vserver 2013-09-26 06:20:39 wwrong win sorry 2013-09-26 09:58:15 clandmeter: do you have any opinion on ruby-gems? 2013-09-26 09:58:23 ok if we just drop that subpackage? 2013-09-26 09:58:34 and ship it with 'ruby' and 'ruby-libs' 2013-09-26 09:59:28 didnt we that already fix in edge? 2013-09-26 09:59:33 nope 2013-09-26 09:59:50 i think we talked about it earlier 2013-09-26 09:59:56 apk add ruby doesnt add gems? 2013-09-26 10:00:02 nope 2013-09-26 10:00:15 ok then i only thought we fixed it 2013-09-26 10:00:26 we did talk about it 2013-09-26 10:00:38 i remember taht we talked about it 2013-09-26 10:00:44 but i dont think it got fixed 2013-09-26 10:00:44 Whats best? 2013-09-26 10:00:51 just drop the subpkg? 2013-09-26 10:00:53 i dont know 2013-09-26 10:00:57 thats the simplest for sure 2013-09-26 10:01:39 sounds ok 2013-09-26 10:01:45 ok 2013-09-26 10:01:55 i'll do it for edge 2013-09-26 10:02:05 there are lots of changes in edge 2013-09-26 10:02:08 but its on the tracker 2013-09-26 10:02:13 for ruby i mean 2013-09-26 10:02:17 ok 2013-09-26 10:47:33 Okay, maybe I'm derptastic here, but.. I'm trying to use abuild 2013-09-26 10:48:03 First it whines that I'm root. Errors out immediately with "Don't use abuild as root". 2013-09-26 10:48:17 So I make a new user and.. /usr/bin/abuild: line 1: can't create /var/cache/distfiles/syslinux-6.01.tar.xz.lock: Permission denied 2013-09-26 10:48:31 you need to be in abuild group iirc 2013-09-26 10:48:43 Aah, there is a group? I didn't see that on the wiki. 2013-09-26 10:48:54 abuild should complain about it 2013-09-26 10:49:53 Ah ha! Adding the user account to the abuild group did work! 2013-09-26 10:50:07 abuild didn't complain, nope, it just said "Permission denied" as I showed ^ 2013-09-26 10:50:17 But thank you, barthalion, that did resolve the issue 2013-09-26 10:52:04 np 2013-09-26 10:52:15 well, it should, I remember it did here 2013-09-26 10:52:25 but I'm using edge 2013-09-26 10:52:48 you should take a look at abuild-keygen too 2013-09-26 10:55:28 I'm on edge, too, trying to see what I can do to bring syslinux up to 6.x which was just released last month, brings native EFI :) 2013-09-26 10:55:47 ask ncopa about progress 2013-09-26 10:55:54 I'm sure we've spoken about it 2013-09-26 10:56:06 but I don't remember the decision 2013-09-26 10:56:31 hehe, one issue at a time.. he is working on a bug I found in the ruby package right now 2013-09-26 10:59:58 hi 2013-09-26 11:00:51 no i was on lunch :) 2013-09-26 11:01:20 :D 2013-09-26 11:05:23 ncopa: well, it was more that I don't want to create more work for you, I feel it better to let one bug be settled before opening another.. besides, this isn't so much "bug" as "feature request" :) 2013-09-26 11:17:51 barthalion: thansk for vlc-2.1.0 :) 2013-09-26 11:18:07 np :) 2013-09-26 11:18:49 I think we can push ffmpeg 2.0 2013-09-26 11:19:21 ah, yes 2013-09-26 11:19:25 anyway 2013-09-26 11:19:34 iirc freerdp requires patching 2013-09-26 11:19:52 yup 2013-09-26 11:19:53 https://projects.archlinux.org/svntogit/community.git/tree/trunk/ffmpeg2.0.patch?h=packages/freerdp 2013-09-26 11:20:45 i think my armv6 bootstrap suddenly started worked 2013-09-26 11:21:10 could be this http://git.alpinelinux.org/cgit/aports/commit/?id=c1a172eeffd00a81e1f6c3ebc6f92d5edafc8a4f 2013-09-26 11:23:54 barthalion, fixing build != fixing vlc. the patches are there for a reason 2013-09-26 11:24:14 you even deleted the trigger. that breaks the plugin cache. 2013-09-26 11:24:22 I don't want to block buildbots 2013-09-26 11:24:29 I'm reworking patches right now 2013-09-26 11:25:21 ok 2013-09-26 11:25:35 the first built vlc package will just be then utterly broken 2013-09-26 11:28:49 it seems that everything was merged anyway 2013-09-26 11:29:49 it still failed to build without patches 2013-09-26 12:24:47 ncopa: build64-2-3 offline? seems like build-2-6 and build64-2-6 offline too 2013-09-26 12:28:41 algitbot: build master 2013-09-26 12:29:36 jbilyk: there are no build-2-6. they get messages from git hooks and algitbot 2013-09-26 12:30:36 oh right 2013-09-26 12:30:39 thx ncopa :) 2013-09-26 12:49:09 ncopa: how about ruby-bin pkg ? 2013-09-26 12:49:39 and rub pkg would be meta for ruby-bin+ruby-gems... 2013-09-26 12:50:39 ruby binary has a command-line option --disable-gems 2013-09-26 12:50:58 just a thought 2013-09-26 12:53:16 so users add --disable-gems when gems not installed 2013-09-26 13:01:41 ruby-bin? 2013-09-26 13:02:31 wow, interesting 2013-09-26 13:02:43 ruby --disable-gems does work 2013-09-26 13:05:32 I started to fiddle around ruby and discovered too :) 2013-09-26 13:07:22 i just removed ruby-gems subpackage 2013-09-26 13:07:27 so possibly users just wanting the bin could add --disable-gems as alias and continue to run ruby scripts 2013-09-26 13:07:29 it was the simplest 2013-09-26 13:07:42 now they cannot do that anymore... 2013-09-26 13:07:44 noticed, sorry could not get to logs a bit earlier 2013-09-26 13:08:24 ruby-gems was 1MB 2013-09-26 13:08:35 ruby-libs was 11MB 2013-09-26 13:08:39 now its 12MB 2013-09-26 13:08:44 or something like that 2013-09-26 13:18:07 algitbot: build master 2013-09-26 13:26:41 so at least it compiles 2013-09-26 13:26:52 what is failing? 2013-09-26 13:27:02 ld 2013-09-26 13:27:29 sound like underlinking 2013-09-26 13:27:38 do we have fminf in uclibc? 2013-09-26 13:27:59 nope 2013-09-26 13:28:01 we dont 2013-09-26 13:28:32 we have fminl 2013-09-26 13:36:00 i suppose we could just use fmin() 2013-09-26 13:36:19 http://sprunge.us/ZcdF 2013-09-26 13:37:23 or just define it ourselves... 2013-09-26 13:37:28 http://git.alpinelinux.org/cgit/aports/tree/main/wine/uclibc-fmaxf-fminf.patch?id=c9b491b6099eec02a835ffd05539b5c783c6c43a 2013-09-26 13:55:16 i suppose a similar patch should be sent upstream 2013-09-26 13:55:27 but i think it there should be a test in configure.ac 2013-09-26 13:55:50 and #if !defined(HAVE_FMINF) 2013-09-26 13:56:11 and finally it should probably be defined as inline 2013-09-26 14:02:39 \o/ 2013-09-26 14:25:19 nice! 2013-09-26 14:52:44 vlc works :) 2013-09-26 16:08:17 (3/13) Upgrading vlc (2.0.8a-r0 -> 2.1.0-r0) 2013-09-26 16:08:17 ERROR: vlc-2.1.0-r0: trying to overwrite usr/lib/vlc/plugins/codec/libx264_plugin.so owned by vlc-xorg-2.0.8a-r0. 2013-09-26 16:08:33 (8/13) Upgrading ruby (2.0.0_p247-r1 -> 2.0.0_p247-r2) 2013-09-26 16:08:33 ERROR: ruby-2.0.0_p247-r2: trying to overwrite usr/bin/gem owned by rubygems-1.3.5-r5. 2013-09-27 12:24:40 tdtrask: are u there? 2013-09-27 12:37:27 alacerda: hello 2013-09-27 12:38:36 were with you who I talked some months ago about ACF devel? ( ok if you don't remember ) 2013-09-27 12:39:09 probably :) 2013-09-27 12:39:30 but, no, I don't remember 2013-09-27 12:39:47 hehe.. I was following wiki.al.o but you told me that maybe that was not updated. 2013-09-27 12:41:03 ACTION still hasn't updated the docs 2013-09-27 12:41:28 they're good for background info 2013-09-27 12:41:39 but the code is the real documentation :/ 2013-09-27 12:42:07 This is the point. Even if you didnt updated it. Do you have any in-off doc that need to be tested? 2013-09-27 12:42:44 I think reading the code i can get it as you say... but, slower than if i read documentation 2013-09-27 12:42:58 If you have something that need to be tested. 0/ 2013-09-27 12:43:18 on how to devel apps for ACF 2013-09-27 12:51:53 sorry, was afk, now back with food :) 2013-09-27 12:52:21 thanks for the offer 2013-09-27 12:53:02 ACTION is trying to squeeze in some changes before alpine 2.7 is released next month 2013-09-27 12:53:06 kind of swamped now 2013-09-27 12:54:20 any aspect in particular you are interested in? 2013-09-27 12:57:22 tdtrask: i'm just back to my first steps on ACF. But if what we have is there on wiki.al.o, no problem 2013-09-27 12:57:24 i can get by 2013-09-27 12:57:28 thanks 2013-09-27 12:59:16 ACTION will do what he can 2013-09-27 12:59:37 ;) 2013-09-27 14:44:15 hrmpf 2013-09-27 14:44:19 seems that sabotage cares only about musl 2013-09-27 14:44:25 and breaks uclibc/glibc side often 2013-09-27 14:44:49 need to be more careful with them 2013-09-27 14:53:40 back later today to fix any possible other breakage 2013-09-27 15:03:52 hmm 2013-09-27 15:06:01 fabled: ./hdb.h:44:22: fatal error: hdb_asn1.h: No such file or directory 2013-09-27 15:07:05 from where? 2013-09-27 15:07:47 http://dev.alpinelinux.org/buildlogs/build64-edge.1380293025.log 2013-09-27 15:07:51 oh 2013-09-27 15:08:05 why did it succeed on build, but failed on build-64 ? 2013-09-27 15:08:16 sounds weird 2013-09-27 15:08:21 algitbot: retry master 2013-09-27 15:08:25 algitbot: build master 2013-09-27 15:12:49 we have now musl-x86 builder on the build vserver host 2013-09-27 15:13:50 have built about 300 of the sources in edge/main 2013-09-27 15:16:32 oh, uclibc has patchwork now 2013-09-27 15:16:41 and they didn't even set it up! 2013-09-27 15:16:45 someone else set it up 2013-09-27 15:17:39 yeah 2013-09-27 15:17:45 let's hope the start use it 2013-09-27 15:18:39 would be nice, but hey, there's musl now :P 2013-09-27 15:19:54 well, soon anyway :) 2013-09-29 07:36:00 barthalion: around? 2013-09-29 07:45:04 yup 2013-09-29 07:45:45 clandmeter: ↑ 2013-09-29 07:50:59 barthalion: hi 2013-09-29 07:51:40 \o 2013-09-29 07:51:44 should we still wait for xcache? 2013-09-29 07:51:57 i see arch is using svn for xcache 2013-09-29 07:52:13 yes, svn version works fine with 5.5 2013-09-29 07:52:40 is that a special rev they are using? 2013-09-29 07:52:58 not really, maintainer just bump it from time to time 2013-09-29 07:53:08 currently it is 1394 2013-09-29 07:53:37 ok, then ill upgrade it this week if somebody else didnt beat me to it 2013-09-29 07:57:10 barthalion: is there still a reason to use xcache with 5.5? 2013-09-29 07:57:32 possibly no 2013-09-29 07:58:00 there is apcu and built-in opcache from zend 2013-09-29 07:58:13 (apcu is external one) 2013-09-29 07:58:58 yes 2013-09-29 08:00:03 I guess most of them want to have some kind of "easier" upgrade path... 2013-09-29 08:00:15 yeah, that's the main point I guess 2013-09-29 08:00:41 and not everything works fine with built-in module yet 2013-09-29 08:00:49 I had some issues with phpbb3 2013-09-29 08:02:36 ok good to know 2013-09-29 08:03:01 i need to set one up for a client 2013-09-29 08:04:04 I mean, everything is fine until you want to define cache method 2013-09-29 08:04:28 only plaintext works, not sure if zend module works on its own without any settings from app side 2013-09-29 08:04:54 I wasn't interested enough, it wasn't for a client :p 2013-09-29 08:15:14 anybody know how to get owncloud/alpine and utf8 to play nice? 2013-09-29 09:16:45 fabled: getting close? 2013-09-29 09:17:42 still have few issues on todo list - but yes, things are starting look a lot better now. 2013-09-29 09:17:51 i think i got the armv6 toolchain working too. 2013-09-29 09:18:04 the musl changes are based on x86 builder though 2013-09-29 09:18:19 it has now built 947 of 1661 packages in main 2013-09-29 09:18:29 what does that mean? 2013-09-29 09:18:35 regarding x86 builder 2013-09-29 09:19:06 you dont build with wandboard? 2013-09-29 09:21:07 i want to fix build against musl first on x86 since it's faster 2013-09-29 09:21:17 i think i could trigger a build on wandboard too 2013-09-29 09:21:23 just want lxc working first 2013-09-29 09:21:29 so i get proper build host name 2013-09-29 09:23:36 you will have a complete musl x86 toolchain and packages later? 2013-09-29 09:23:54 i have complete complete musl x86 toolchain, and builder vserver setup 2013-09-29 09:24:03 and i have almost 60% of aports/main built for that 2013-09-29 09:24:27 ok 2013-09-29 09:24:52 i have also complete musl arm toolchain for armv6 and armv7, and bootstrap images. + wandboard that boots on them 2013-09-29 09:25:23 you want to replace x86 uclibc or this is just to support the arm one? 2013-09-29 09:25:28 both 2013-09-29 09:25:38 eventually we want x86 uclibc replaced with musl 2013-09-29 09:25:56 but it also helps doing the arm build - the x86 builder is way faster 2013-09-29 09:26:06 I guess you want to have them side by side initially? 2013-09-29 09:26:17 possibly 2013-09-29 10:11:49 huh 2013-09-29 10:12:55 algitbot: build master 2013-09-29 10:13:47 I'll take a look later 2013-09-29 14:10:07 please git pull git://git.alpinelinux.org/fab/aports 2013-09-29 14:10:50 now it's nikola, httpie, the deps, and some updates 2013-09-29 14:11:11 thanks 2013-09-29 14:24:15 >>> ERROR: py-doit: doit-0.23.0.tar.gz is missing in checksums 2013-09-29 14:24:21 fabian_a, ^^ 2013-09-29 14:26:02 good start :-( let me fix that 2013-09-29 14:30:37 fabled: checksums are now up-to-date 2013-09-29 14:42:57 >>> ERROR: py-pillow: Pillow-2.1.0.zip is missing in checksums 2013-09-29 14:43:01 fabian_a, ^^ 2013-09-29 14:43:13 can you check the others too? 2013-09-29 14:45:11 fabled: i'm on it 2013-09-29 14:56:57 fabled: two more errors fixed 2013-09-29 15:13:15 fabled: thanks 2013-09-29 17:00:35 Anyone familiar with when Alpine refuses to send an email on an SMTP server and returns error 550 5.7.0 and says that my from address isn't one of my address? 2013-09-29 17:15:21 Anyone even here? 2013-09-29 17:31:57 this is not help 2013-09-29 17:32:01 for the alpine mail client 2013-09-29 17:32:08 sorry 2013-09-29 17:38:53 whoops, wrong channel.. argh, bad morning, sorry 2013-09-30 06:21:01 ugh 2013-09-30 06:21:05 ruby-gems.... 2013-09-30 06:21:19 we have a ton of ruby-* packages that depends on ruby-gems 2013-09-30 07:39:01 #2261 2013-09-30 07:39:11 bug 2261 2013-09-30 09:35:22 ncopa: you've been busy ;-) 2013-09-30 11:20:05 anyone has succesfully booted alpine on a macbook? 2013-09-30 11:28:16 please git pull git://git.alpinelinux.org/fab/aports 2013-09-30 11:29:17 this request contains willie, the needed deps, and some sphinx modules 2013-09-30 11:30:38 ACTION just thinking loud...fcolista, aren't recent macbooks built with intel hardware? 2013-09-30 11:31:03 yep, the "problem" is EFI boot 2013-09-30 11:47:10 fabian_a, yes 2013-09-30 11:47:26 but UEFI boot is not so immediate to boot 2013-09-30 11:47:44 http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/arch/x86/boot/compressed/eboot.c?id=f8b8404337de4e2466e2e1139ea68b1f8295974f 2013-09-30 11:48:05 syslinux 6 supports efi 2013-09-30 11:48:44 https://bugs.archlinux.org/task/33745 <--- at the end of the page 2013-09-30 11:49:05 i've created an usb thumb with alpine64 2013-09-30 11:49:07 gpt 2013-09-30 11:49:10 partition 2013-09-30 11:57:10 fcolista, funny you should mention that 2013-09-30 11:57:39 fcolista, I am right now - literally, in another terminal - attempting to make a syslinux 6 aport and make EFI boot work with Alpine on a MacBook Pro 2013-09-30 11:57:52 fcolista, will keep you posted on my progress if you would like to stick around! :) 2013-09-30 11:58:14 lol! 2013-09-30 11:58:22 foxkit, nice to know 2013-09-30 12:02:42 alpine kernel config come withoud CONFIG_EFI enabled 2013-09-30 12:04:35 don't need CONFIG_EFI 2013-09-30 12:04:39 all we need is syslinux with EFI 2013-09-30 12:04:48 I have this already working with Fedora and Gentoo ^^ 2013-09-30 12:04:57 k 2013-09-30 12:07:29 hrm, actually.. thinking it's about time to snooze 2013-09-30 12:07:54 fcolista, you think you'll be here in maybe 12 hours or so? probably be sleeping for 4 then 8 hours of work 2013-09-30 12:08:03 i'll be here 2013-09-30 12:08:14 oh 2013-09-30 12:08:14 ok, see you then. 2013-09-30 12:08:19 12 hours 2013-09-30 12:08:20 no 2013-09-30 12:08:24 :P 2013-09-30 12:08:24 it's 14:00 now in ITA 2013-09-30 12:08:35 aah, it's 7:08 in central US 2013-09-30 12:08:38 in 12 hours i'm sleeping :) 2013-09-30 12:08:48 I have a weird schedule, I know 2013-09-30 12:09:04 okay well I will be here all night/morning in my time, so tomorrow perhaps I can have something for you to test with 2013-09-30 12:09:14 always nice to have someone else to help test to make sure it works on more than one laptop! 2013-09-30 12:09:31 yep 2013-09-30 12:11:37 night o/ 2013-09-30 12:11:44 _0/ 2013-09-30 12:59:29 ncopa: could you have a look at netatop at atop site, seems nice 2013-09-30 13:07:57 ncopa: you around? 2013-09-30 13:08:40 I recalled pingu compile issue on debian - "strlcpy" not found 2013-09-30 13:09:13 can "replace" pkg be used in such issues (from samba.org) ? 2013-09-30 13:09:25 to solve the problem 2013-09-30 13:13:24 http://www.irclogger.com/.alpine-linux/2013-08-21#1377089844 2013-09-30 13:26:18 vkrishn: hi 2013-09-30 13:26:20 im here 2013-09-30 13:26:23 but slightly busy 2013-09-30 13:30:11 hmm, noticed lengthy gem issues 2013-09-30 13:34:02 vkrishn> can "replace" pkg be used in such issues (from samba.org) ? 2013-09-30 13:34:03 ? 2013-09-30 13:34:06 i dont understand... 2013-09-30 13:35:59 this was the fix for strlcpy in pingu: https://github.com/ncopa/pingu/commit/7a7b0bd1a4182201a5006c31ab5f1a4ef5849ee6 2013-09-30 13:44:46 asking in general, was checking http://rsync.samba.org/ftp/unpacked/replace/ and recalled the pingu issue 2013-09-30 13:48:37 ah 2013-09-30 13:48:38 yes 2013-09-30 13:48:47 and there is also gnulib 2013-09-30 13:59:33 ping tdtrask 2013-09-30 13:59:41 morning you and all 2013-09-30 13:59:48 I'm trying this example: http://wiki.alpinelinux.org/wiki/ACF_mvc.lua_example 2013-09-30 14:00:44 and when I try to update hostname from CLI (for while) i recieve: 2013-09-30 14:00:51 lua: (error object is not a string) 2013-09-30 14:01:30 command line was: lua helloworld.lua update newname 2013-09-30 14:02:00 I'm trying Step 5 on the example 2013-09-30 14:35:19 alacerda: hi 2013-09-30 14:35:35 :) 2013-09-30 14:35:37 hi 2013-09-30 14:35:45 any idea? 2013-09-30 14:36:21 fabled: ^^ 2013-09-30 14:36:55 bah 2013-09-30 14:37:09 will fix 2013-09-30 14:37:41 alacerda: that doc hasn't been updated in 5 years :( 2013-09-30 14:38:02 ACF has changed significantly 2013-09-30 14:38:28 But as i can understand untill step 5 should still work 2013-09-30 14:38:47 am i wrong? 2013-09-30 14:40:47 ACTION doesn't know what "step 5" refers to 2013-09-30 14:41:09 hehe... I mean: 2013-09-30 14:41:14 The topic "5. Test the application:" 2013-09-30 14:41:27 sorry 2013-09-30 14:41:31 ok 2013-09-30 14:42:38 ACTION is surprised that step one "svn" works 2013-09-30 14:42:43 since we've been using git for years 2013-09-30 14:42:45 I've followed every steps above this topic. Just replace svn export svn://svn.alpinelinux.org/acf/core/trunk/www/cgi-bin/mvc.lua 2013-09-30 14:42:54 :) 2013-09-30 14:42:59 haha 2013-09-30 14:43:01 opa 2013-09-30 14:43:03 oops 2013-09-30 14:43:26 I did: wget http://git.alpinelinux.org/cgit/acf-core/tree/lua/mvc.lua 2013-09-30 14:43:35 instead of svn 2013-09-30 14:46:31 ACTION is doing a quick rewrite of the doc for current standards of ACF 2013-09-30 14:47:06 ACTION is very thankful 2013-09-30 15:04:35 clandmeter, the latest build failure is due to opensc/pcsc-lite 2013-09-30 15:05:16 fabled: log? 2013-09-30 15:05:26 oh i see 2013-09-30 15:05:56 mv: can't rename '/var/cache/distfiles/pcsc-lite-1.8.8.tar.bz2.part': No such file or directory 2013-09-30 15:05:56 >>> ERROR: pcsc-lite: abuildindex failed 2013-09-30 15:06:04 oh 2013-09-30 15:06:05 also 2013-09-30 15:06:07 abuild: /home/buildozer/aports/testing/bugzilla/APKBUILD: line 39: a: not found 2013-09-30 15:06:49 thats not mine afaik 2013-09-30 15:06:50 fabian_a, fcolista: ^^ 2013-09-30 15:09:27 what? 2013-09-30 15:10:44 bugzilla failed 2013-09-30 15:10:55 it's your aport - though fabian_a had the latest change to it 2013-09-30 15:11:06 k 2013-09-30 15:11:12 just looking at 2013-09-30 15:11:32 there an "a" at the end of APKBUILD 2013-09-30 15:12:47 fabian_a, i think that came from your commit 2013-09-30 15:13:12 i've some unstaged changes because of new aports in testing 2013-09-30 15:13:21 cannot pull --rebase in order to push the change 2013-09-30 15:13:39 (or there's a way to do it?) 2013-09-30 15:16:21 git stash ; git pull --rebase ; git stash apply 2013-09-30 15:17:06 can i push after pull --rebase and then stash apply? 2013-09-30 15:18:05 yes 2013-09-30 15:21:15 alacerda: please review the beginning of the doc 2013-09-30 15:22:07 it's significantly different :) 2013-09-30 16:00:22 tdtrask: thanks. I gonna try. 2013-09-30 16:00:34 sorry for delay, I was in lunch time 2013-09-30 16:21:26 tdtrask: same error that I was facing before. 2013-09-30 16:21:32 look: http://pastebin.com/xPRFd70s 2013-09-30 16:25:02 alacerda: I'm slowly working my way through the doc 2013-09-30 16:25:19 the custom MVC application still needs a rewrite 2013-09-30 16:26:03 ahw, you said "alacerda: please review the *beginning* of the doc"... sorry my bad! 2013-09-30 16:26:55 np 2013-09-30 16:27:09 ACTION is running out and will try to update more later 2013-09-30 16:30:08 tdtrask: thanks 2013-09-30 20:13:22 hello guys 2013-09-30 20:13:53 I just want to tell you something, I have seen that you want to migrate to musl-libs - don't do it! Why - 2013-09-30 20:14:37 well, why noy 2013-09-30 20:14:41 *not 2013-09-30 20:14:45 Right now Alpine linux is the only one which uses Uclibc and has integrated gcc compiler, busybox already excluded gcc from target support 2013-09-30 20:15:40 and for all developers using Uclibc for their projects (I use it to compile my static prduction software) Alpine linux is a savior 2013-09-30 20:15:51 you should bring your concerns on alpine-devel 2013-09-30 20:16:08 anyway, you can make a crosscompiler on glibc or musl system 2013-09-30 20:16:42 Why Uclibc and not musl - because musl does not work on non-linux boxes, for example on Freebsd under linux emulation, but uClibc do! 2013-09-30 20:17:13 musl uses different proc/ filesystem hacks which cannot be easily emulated 2013-09-30 20:18:13 well, that is all I wanted to say :) and thank you for your project! 2013-09-30 20:18:44 barthalion, I am on alpine-devel, what about you ? :P 2013-09-30 20:19:37 oh, I meant mailing list 2013-09-30 20:20:20 barthalion, I can't do it, write a few words there if you can 2013-09-30 20:20:39 I'm not really involved into migration stuff 2013-09-30 20:20:58 we have already made the decision to switch to musl, sorry 2013-09-30 20:21:00 I just keep packages up to date 2013-09-30 20:21:44 supporting uclibc consumes too much project resources 2013-09-30 20:21:54 ah, okay 2013-09-30 20:22:04 there is a reason 'alpine linux is a savior' for uclibc-based development 2013-09-30 20:22:12 yes, uClibs is very slow to correct bugs, I already submited two 2013-09-30 20:22:13 that is because we take uclibc 2013-09-30 20:22:17 and fix it 2013-09-30 20:22:19 and we have 2013-09-30 20:22:22 30+ patches 2013-09-30 20:22:31 that will never ever be merged to be blunt 2013-09-30 20:22:54 kaniini, the main reason is - under emulation for example in Freebsd, programs compilled with uClibs work, but musl don't 2013-09-30 20:23:19 this sounds like a generally unsupported usecase 2013-09-30 20:23:23 kaniini, and that is the only reason why I use these small static libraries 2013-09-30 20:25:29 well, at least I can use old Alpine releases ) 2013-09-30 20:27:54 fabled: sorry, i missed that yesterday 2013-09-30 20:29:45 diet-libc almost dead, uClibc development is very slow, musl active but not usable for me, holy ghost ... 2013-09-30 20:29:59 thanks for fixing that 2013-09-30 20:48:06 to be honest, he should consider talking to dalias about his problem with musl; i am sure that he will find a solution 2013-09-30 20:48:35 the main motive for switching from uclibc to musl was the stark difference in the amount of concern upstream has for realworld issues