2015-02-01 00:32:52 does building an custom iso only work as root? 2015-02-01 10:40:33 you can probably use fakeroot 2015-02-01 17:00:51 I'm getting linking errors when trying to compile libs for dnssec: https://0.jmt.gr/?244967bcf8f8e628#SVcKjiGGIcX/EnQZha73QlFIp9k+6miF+tR8gx4SzbA= 2015-02-01 17:01:22 the gcc call has -lssl 2015-02-01 17:01:58 the libssl at least has matching strings in it: strings /usr/lib/libssl.so.1.0.0|grep EVP_sha1 <- true 2015-02-01 17:03:54 ldd /home/jomat/aports/testing/dnssec-tools/src/dnssec-tools-2.1/validator/libsres/.libs/libsres.so tells me the same.... what am I doing wrong? :-) 2015-02-01 17:04:20 ldd: https://0.jmt.gr/?ef4523dd1810aeab#TzgFiwEwQYQ7US7nDmZLPw2BafABa0+7wEl1XOfoLi8= 2015-02-01 18:49:28 @jomat: - it builds ok for me on 64 bit - I've put the APKBUILD on https://0.jmt.gr/?244967bcf8f8e628#SVcKjiGGIcX/EnQZha73QlFIp9k+6miF+tR8gx4SzbA= 2015-02-01 18:49:57 BitL0G1c: Thx alot! 2015-02-01 18:50:37 BitL0G1c: But does that also build the libs? 2015-02-01 18:51:09 you can sometimes get build errors if you have not recently done an 'apk update && apk upgrade' 2015-02-01 18:51:40 My systems are updated 2015-02-01 18:51:58 builds ok on 32 bit too - it builds dnssec-tools-dev-2.1-r0.apk & dnssec-tools-2.1-r0.apk 2015-02-01 18:52:29 BitL0G1c: But just the binaries? 2015-02-01 18:52:50 the binary & the headers 2015-02-01 18:52:51 I need the libs (i. e. change --without-validator to --with-validator) 2015-02-01 18:56:03 gives Makefile:105: recipe for target 'perlmake' failed 2015-02-01 18:57:50 https://0.jmt.gr/?244967bcf8f8e628#SVcKjiGGIcX/EnQZha73QlFIp9k+6miF+tR8gx4SzbA=' 2015-02-01 19:00:34 Oh... on a 32 bit system? 2015-02-01 19:01:47 and you have musl-dev installed? 2015-02-01 19:14:51 gives the same on on 32 & 64 - it needs a CFLAGS=”-D_LARGEFILE64_SOURCE=1″ 2015-02-01 19:15:16 but with that it doesn't build 2015-02-01 19:16:25 btw... what I think I need it for: https://github.com/irssi/irssi/commit/d826896f74925f2e77536d69a#diff-3b3a6ec97232deb43dc14319a73872c1R631 2015-02-01 19:20:07 ok got it to build 2015-02-01 19:20:40 before configure add: export CFLAGS="$CFLAGS -D_LARGEFILE64_SOURCE=1" 2015-02-01 19:24:48 & add to package(): rm -f "$pkgdir"/usr/lib/*.la 2015-02-01 19:27:36 And then the resulting package contains a libsres.so and libval-threads.so ? 2015-02-01 19:30:53 BitL0G1c: I think I have the wrong version of libssl 2015-02-01 19:31:03 BitL0G1c: Can you show me the output of apk policy libssl1.0 2015-02-01 19:31:49 that's mine: https://0.jmt.gr/?b7942454863baa23#GQqlaHV9WUcHHYofvR620Sq/ycgWtQhZ105bUe43tgo= 2015-02-01 19:32:43 1.0.1l-r0: 2015-02-01 19:33:30 BitL0G1c: rm -rf is uneccessary. http://git.alpinelinux.org/cgit/abuild/commit/?id=fa204f85578296135d19987b55ddeb6622ff68a8 2015-02-01 19:35:02 without it - the build fails 2015-02-01 19:36:18 jomat: yes those libs are in the build https://0.jmt.gr/?244967bcf8f8e628#SVcKjiGGIcX/EnQZha73QlFIp9k+6miF+tR8gx4SzbA= 2015-02-01 19:36:41 thx alot! 2015-02-01 19:37:16 ohno.... I think I made a mistake: sudo: apk: command not found 2015-02-01 19:37:20 :-D 2015-02-01 19:38:34 I will submit a patch 2015-02-01 19:39:42 In the meantime i try to repair my apk.... oh gosh, it also killed the firefox 2015-02-01 19:40:45 scadu: - see http://hastebin.com/ojopukirev.lua 2015-02-01 19:45:23 scadu: there was no new release of abuild yet 2015-02-01 19:45:29 so atm yes, it's still needed 2015-02-01 19:47:15 oh right. 2015-02-01 19:51:39 is 'patchelf' in alpine ? 2015-02-01 19:53:48 you want to remove insecure rpath? 2015-02-01 19:53:54 rpath is in main for that 2015-02-01 19:53:55 yes 2015-02-01 19:54:06 ok 2015-02-01 19:54:23 *chrpath* 2015-02-01 19:57:17 ok 2015-02-01 20:00:22 i think there is a problem with disk installs http://hastebin.com/podihexecu.vbs 2015-02-01 20:00:47 that was from a setup-disk -m sys 2015-02-01 20:01:04 catch ncopa with that 2015-02-01 20:01:16 things like that are beyond my alpine understanding :p 2015-02-01 20:02:01 me too 2015-02-01 20:03:04 will raise an issue on the bug tracker 2015-02-01 20:03:10 /* Make sure the file is part of the cached directory tree */ 2015-02-01 20:03:21 I know some of these words 2015-02-01 20:04:03 but apk-tools source code is beyond me as well, so yeah, report it 2015-02-01 20:04:13 fabled and ncopa hold all the answers 2015-02-01 20:09:30 I'm almost ready to add a patch to setup-disk to add support for LUKS encrypted installs. I've also written a terminal partition editor 2015-02-01 21:28:52 hi. I'm trying to create an account on the alpine linux wiki, but I don't know what to provide to the "Are you a spammer?" question 2015-02-01 22:19:38 "no" should work 2015-02-01 22:19:41 :p 2015-02-01 22:23:03 barthalion: 23:00:52 -!- alpine_tester [~adrianfi@188.24.123.27] has quit [Quit: Leaving] 2015-02-01 22:28:56 barthalion: no doesn't work for me :-/ 2015-02-02 01:33:40 Can linux-grsec-3.14.30-r0 be rebuilt please for 32 bit - the archive is corrupt http://bugs.alpinelinux.org/issues/3847 (64 bit is unaffected) 2015-02-02 05:48:23 scadu: filter joins and parts and make your life easier 2015-02-02 06:45:18 morning 2015-02-02 06:45:31 cool raspberry pi 2 unveiled, that is quad core armv7 2015-02-02 06:46:01 cost? 2015-02-02 06:47:29 same as B+ 2015-02-02 06:47:37 http://www.theregister.co.uk/2015/02/02/raspberry_pi_model_2/ 2015-02-02 06:47:54 $35 2015-02-02 06:49:55 nice 2015-02-02 06:50:10 i wonder what nic it has 2015-02-02 06:52:28 hmm 2015-02-02 07:02:29 i wonder why raspberrypi.org does not mention it 2015-02-02 07:05:57 the other article i found says it's targeted for 2017 2015-02-02 07:10:01 not so nice then 2015-02-02 07:10:27 yeah. that's a bummer. 2015-02-02 07:10:42 sounds more plausible still since B+ just come out few months ago 2015-02-02 07:11:04 btw. is anyone working on libreoffice packaging? 2015-02-02 07:11:37 I don't think so, or he wasn't too loud about it 2015-02-02 07:58:28 http://allegro.pl/pamiec-ram-2gb-pc3-10600-i5044358035.html brać, nie brać 2015-02-02 07:59:42 ffu, wrong window. 2015-02-02 10:07:17 fabled: we have a new competitor for rpi v2 ;-) http://blogs.windows.com/buildingapps/2015/02/02/windows-10-coming-to-raspberry-pi-2/ 2015-02-02 10:47:15 hups 2015-02-02 10:47:26 i managed to specify wrong CVE number in 4 commits... 2015-02-02 10:47:29 5 commits 2015-02-02 14:35:02 fwiw, i'm working on libreoffice currently 2015-02-02 16:17:57 fabled: \o/ 2015-02-02 17:22:12 > | 2015-02-02 17:38:50 > | 2015-02-03 08:05:49 ncopa: do you have a second? I'm trying to create an iso with custom pacakages added, not fetched and put in /media/apks. Do you have any tips? 2015-02-03 08:09:40 morning 2015-02-03 08:10:01 ScrumpyJack: whats the name of your iso profile? 2015-02-03 08:10:16 if you have a alpine-scrumpyjack.conf.mk 2015-02-03 08:10:33 then you need an alpine-scrupyjack.packages file with the list of packages you want 2015-02-03 08:10:50 finally you build with: fakeroot make PROFILE=alpine-scrumpyjack 2015-02-03 08:17:23 maybe i'm doing something wrong, because alpine-scrupyjack.packages doesn't seem to install the packages, it just creates a repo in the iso with that list 2015-02-03 08:17:59 and if you look at the Makefile, it suggests that it only does an apk fetch on your list of custom pacakges 2015-02-03 08:22:34 do i need to get them in the apkvol? 2015-02-03 08:22:46 and in world? 2015-02-03 08:40:07 ah 2015-02-03 08:40:16 yes, it will just create a repo with those packages 2015-02-03 08:40:27 the iso will not install any packages at all at boot 2015-02-03 08:40:36 only as few as possible to get it started 2015-02-03 08:40:44 you will have to provide an apkovl for that 2015-02-03 09:23:08 whee 2015-02-03 09:23:10 libreoffice built 2015-02-03 09:23:12 and runs 2015-02-03 09:23:14 \o/ 2015-02-03 09:23:17 congrats! 2015-02-03 09:23:19 now to just package it 2015-02-03 09:23:28 and fix all the prerequisites 2015-02-03 09:23:29 that is an impressive achievement 2015-02-03 09:23:38 it needed few small patches 2015-02-03 09:23:42 and found an annoying gcc bug 2015-02-03 11:22:47 ncopa: is there better way to getting the cd to boot with pre-installed packages than creating an apkvol? 2015-02-03 11:54:04 hi 2015-02-03 11:55:43 can anyone here define a how alpine build system compare to bitbake/oe? 2015-02-03 11:57:09 fabled: ↑ 2015-02-03 12:00:16 i had been doing some work on enabling yocto to build container images, and i can draw a comparison between bitbake and buildroot as well as portage... 2015-02-03 12:00:36 ScrumpyJack: not that i know 2015-02-03 12:00:49 but i just discovered alpine, and wonder where exactly it stands 2015-02-03 12:01:08 errordeveloper: i havent use bitbake/oe 2015-02-03 12:01:18 but alpine does not do cross-compile 2015-02-03 12:01:50 i suspect bitbake/oe supports crosscompile 2015-02-03 12:01:56 that's right 2015-02-03 12:01:59 ok, i see 2015-02-03 12:02:28 with alpine you would do native compiling of your packages, then to generate the root: apk add --root /tmp/root ... 2015-02-03 12:08:36 does alpine have multi-threaded build queueing etc? 2015-02-03 12:15:19 no 2015-02-03 12:15:45 so autoconf will kill performance for you 2015-02-03 12:16:10 you can ofcourse set make flags to -j24 if you want 2015-02-03 12:16:12 which we do 2015-02-03 12:16:24 but it will not build different packages at same time 2015-02-03 12:18:12 ok, cool 2015-02-03 12:18:28 ncopa: thanks for the input 2015-02-03 12:19:24 i will probably continue hacking on yocto, as i'm quite close to one of my usability targets there... 2015-02-03 12:19:53 it may be of interest to port APK to bitbake thought, for the fun of it 2015-02-03 12:20:31 anyhow, it's great to see that people have interest in minimal container runtime :) 2015-02-03 12:21:18 i certainly would be looking forward to join efforst in some ways and give alpine a spin for some of my daily work ;) 2015-02-03 12:22:43 the big dilema of mine is how to explain to developers that use python/ruby/node.js what's the difference between minimal runtime-only container images and what most folks currently do, where they bundle all build-time deps into their containers... 2015-02-03 12:23:52 understand your dilemma :) 2015-02-03 15:01:37 okay 2015-02-03 15:01:51 libreoffice split still to do 2015-02-03 15:01:57 and then i'll push it 2015-02-03 15:02:09 it seems to at least start, and open .pdf files and basic stuff works 2015-02-03 15:23:37 how nice. libreoffice ships it's own file split lists 2015-02-03 15:23:45 i should have pkg done tomorrow 2015-02-03 17:31:12 ok. that's all libreoffice prerequisites. 2015-02-03 17:31:19 libreoffice to follow tomorrow 2015-02-03 17:31:24 (if all goes well) 2015-02-03 17:38:51 hmmh 2015-02-03 18:26:46 The following NEW packages will be installed: 2015-02-03 18:26:46 libreoffice-base libreoffice-calc libreoffice-common libreoffice-connector-postgres libreoffice-draw libreoffice-gnome 2015-02-03 18:26:46 libreoffice-impress libreoffice-lang-en-us libreoffice-math libreoffice-writer 2015-02-03 18:26:46 After this operation, 265 MiB of additional disk space will be used. 2015-02-03 18:28:19 humm 2015-02-03 18:30:05 fabled: what's weird about it? 2015-02-03 18:30:25 now that it's installedfrom packages, it hangs 2015-02-03 18:30:33 it worked from the build dir 2015-02-03 18:30:34 splitting libreoffice is mostly not worth the candle 2015-02-03 18:31:28 barthalion, libreoffice gives the filelists automatic 2015-02-03 18:31:32 it was simple enough 2015-02-03 18:31:45 until something breaks 2015-02-03 18:32:00 the maintainer of these lists resigned 2015-02-03 18:32:52 and in Arch we had some reports about weird dependencies (draw didn't start without calc) or files in wrong locations 2015-02-03 18:36:47 i see 2015-02-03 18:48:25 ok. libre works now after reboot. apparently i had updated mesa or other stuff that needed X restart 2015-02-03 23:44:22 hi all, where is libdevmapper.h in alpine ? 2015-02-03 23:50:49 ok found it ... 2015-02-04 08:54:40 \o/ 2015-02-04 08:54:48 libreoffice with musl libc... 2015-02-04 08:54:54 let's see if it compiles ;) 2015-02-04 08:55:47 arm has build fails 2015-02-04 08:55:51 webkitgktk 2015-02-04 08:55:55 serf 2015-02-04 08:55:56 firefox 2015-02-04 09:04:39 btw. rpi2 is in store 2015-02-04 09:04:57 seems the other article about 2016 was wrong 2015-02-04 09:05:04 raspberrypi.org is updated too now 2015-02-04 09:05:06 ... 2015-02-04 09:05:19 i wonder if we should change over to armv7 after one or two release cycles 2015-02-04 09:05:54 sounds fine 2015-02-04 09:07:33 you have any experience with thumb2 builds? 2015-02-04 09:08:05 then again, if ncopa gets the additional arm builders up, we might consider doing armv7 build separately 2015-02-04 10:07:15 i'm going on vacation now 2015-02-04 10:07:33 4 weeks 2015-02-04 10:08:41 have fun and don't stay up too long :p 2015-02-04 10:09:48 ncopa: nice 2015-02-04 10:09:53 brazil? 2015-02-04 10:11:29 yup 2015-02-04 10:11:53 ACTION is jealous 2015-02-04 10:12:07 it would be nice if you could try help apply patches from alpine-aports 2015-02-04 10:13:13 alpine-aports? 2015-02-04 10:14:21 the mailing list 2015-02-04 10:14:25 for patches 2015-02-04 10:14:42 sure 2015-02-04 10:15:13 don't worry, we won't break everything 2015-02-04 10:17:28 oh there is a new mailing list 2015-02-04 10:17:37 was that announced somewhere? 2015-02-04 10:18:26 barthalion: don't be so sure ;f 2015-02-04 10:18:46 clandmeter: http://lists.alpinelinux.org/alpine-devel/4622.html 2015-02-04 10:19:43 ncopa, lets drink a coconut water here in brazil :P 2015-02-04 10:21:18 see u 2015-02-04 10:22:14 never received that email. 2015-02-04 10:46:18 got error "/lib/libnih.so: undefined reference to `libintl_gettext'" while building package 2015-02-04 10:46:45 libnih-dev and libnih is installed - gettext also - any ideas ? 2015-02-04 10:56:13 sounds like libintl is being used, but not linked in 2015-02-04 10:56:28 on glibc libintl is part of glibc 2015-02-04 10:56:38 so that's why maybe the code is broke 2015-02-04 10:57:11 you can try a) building without gettext-dev (the dev package for libintl) 2015-02-04 10:57:22 or b) figure out how to add -lintl to LDFLAGS 2015-02-04 11:07:35 fabled: are you running xorg from edge? 2015-02-04 11:07:44 clandmeter, yes 2015-02-04 11:07:45 im getting this error http://sprunge.us/ieUU 2015-02-04 11:08:35 clandmeter, strange. works for me. 2015-02-04 12:57:16 bah. i should have asked the magic scripts from ncopa before he left for vacation 2015-02-04 17:10:44 https://github.com/scadu/aports/commits/master upgraded libfm and polarssl. also, removed Barthalion from maintainer line on his requeest. 2015-02-04 17:22:40 scadu, the commits are mixed 2015-02-04 17:22:52 and i don't think we allow packages in main without maintainer 2015-02-04 17:23:02 could you add yourself as maintainer then? 2015-02-04 17:24:15 fabled: oops, sorry. sure, I'll fix it. 2015-02-04 17:40:16 fabled: fixed 2015-02-04 17:43:31 scadu, pulled, thanks 2015-02-04 17:48:15 trying to build xtrabackup 2.2.8 and I stuck with this: 2015-02-04 17:48:16 /home/builder/xtrabackup/src/percona-xtrabackup-2.2.8/cmd-line-utils/libedit/chartype.h:57:3: error: #error wchar_t must store ISO 10646 characters 2015-02-05 07:55:44 someone has an idea about how to track priorities on openrc boot? In "default" runlevel, how can i determine which come first then another if, e.g., the only "need" is net? 2015-02-05 07:55:53 How OpenRC decides? 2015-02-05 07:59:04 fcolista, you need to specify dependencies using 'need, use, ...' keywords 2015-02-05 07:59:12 if there's no explicit dependency, the order is undefined 2015-02-05 08:00:35 i was scared about that. 2015-02-05 08:00:39 umh. 2015-02-05 08:01:06 So if i have two init with the same "need net", i don't know which one comes first. 2015-02-05 08:01:40 Risk of circular dependencies is quite high if you have several services. 2015-02-05 08:01:54 but, ok, i got the answer. Thx. 2015-02-05 08:02:51 fcolista, there's no easy answer, and the correct order might depend even on install 2015-02-05 08:03:01 you can add those in /etc/conf.d/ 2015-02-05 14:24:37 can I buy "Banana-Pi-M2-quad-core" instead of "wandboard quad" ? 2015-02-05 14:25:26 wandboard kinda difficult to get here 2015-02-05 14:50:29 ok , seems more an alternative for rpi2 2015-02-06 07:25:36 BUUURRN 2015-02-06 07:25:59 somehow I feel that either the build will break or bots are down 2015-02-06 07:30:57 yeah 2015-02-06 07:31:50 fabled: ↑ ? 2015-02-06 07:35:49 ? 2015-02-06 07:35:52 they worked yesterday 2015-02-06 07:37:23 hmm 2015-02-06 07:37:46 seems the lastthing it has done is libreoffice i pushed yesterday 2015-02-06 07:39:04 seems the push notify got lost 2015-02-06 07:39:13 weird 2015-02-06 07:41:18 sounds like the git push hook did not work 2015-02-06 07:50:57 mmm 2015-02-06 07:51:04 sounds like network issue in the mqtt bus 2015-02-06 07:51:20 but that was it. git hook did not talk to mqtt bus, and notify of the new commit 2015-02-06 07:58:23 thanks 2015-02-06 16:47:40 fabled: ping 2015-02-06 18:31:41 what's up with the SSL cert at alpinelinux.org, anyone know? 2015-02-06 18:31:56 clandmeter: ↑ 2015-02-06 18:32:17 what's the problem? 2015-02-06 18:32:21 looks fine to me 2015-02-06 18:33:09 ... woah ... either it was *just* updated, or foul-pkay 2015-02-06 18:38:28 I was just browsing and the Alpine News ended at 3.0.6 release. However, may be false alarm as Firefox complains about SSL cert yet Chromium does not. Probably problem with ca-certs on my end. 2015-02-06 18:39:38 might be, my fx is happy 2015-02-06 18:41:54 I can explain the oddity in the news. :facepalm: I was at forum.alpinelinux.org which seems to be behind the times ;-) 2015-02-06 22:16:42 i have an alpine package for emacs without x11 and other dependencies. https://www.ctrlc.hu/~stef/emacs-nox11-24.99-r0.apk and the package + patches is here: https://www.ctrlc.hu/~stef/emacs-alpine.txz 2015-02-07 10:24:13 what is status of docker image? 2015-02-07 10:28:25 is this latest and official https://github.com/uggedal/docker-alpine 2015-02-07 11:13:44 official, no, but it's surely the best one around 2015-02-07 20:50:18 fabled, are we jumping from kernel 3.14.x to 3.18.x ? :) 2015-02-07 20:50:55 just saw overlay feature in virtgrsec added 2015-02-07 21:01:00 hmm, ok, that depends on grsecurity next stable release 2015-02-07 21:02:20 gotta go 2015-02-07 21:21:36 hm, Arch has linux-grsec 3.18 2015-02-08 08:38:35 barthalion ok thank then ill try it, what was about fake image on docker, is it now removed? 2015-02-08 09:12:32 no idea, didn't check 2015-02-08 22:33:47 huhu? 2015-02-08 22:34:20 cold, isn't it? 2015-02-08 22:35:05 a fair bit yes. 2015-02-08 22:36:36 thing is i managed to get alpine booting and running and all from usb but each time i try to fetch apkindex i get an ioerror 2015-02-08 22:36:46 any clue on that? 2015-02-08 22:37:49 what error exactly? 2015-02-08 22:38:09 some time ago there was reports about 2015-02-08 22:38:14 ehci_hcd module causing issues 2015-02-08 22:38:19 but I guess not in Alpine 2015-02-08 22:38:47 error failure to fetch apkindex.tar.gz ioerror. somesuch 2015-02-08 22:39:35 would be great if you manage to get the exact error 2015-02-08 22:39:45 I'm going to bed for a while, see you in the morning 2015-02-08 22:40:20 well, at least irssi should be in the cache 2015-02-08 22:40:47 gimme a minute. need to reboot for that 2015-02-08 22:40:52 k 2015-02-09 01:00:02 Now that the RPi2 is out, has any thought gone into adding support for it to Alpine? 2015-02-09 01:00:20 (At the very least I believe this involves shipping an ARMv7 kernel) 2015-02-09 01:12:20 XgF: we have support for rpi, which i guess should run also on v2 2015-02-09 01:12:43 clandmeter: An ARMv6 kernel won't run on an ARMv7 device (and vise versa) 2015-02-09 01:13:19 (It could, but typical Linux kernel builds seem to be engineered not to do so. Anyway, it's significantly preferable to ship a kernel per arch major version) 2015-02-09 05:06:24 I want to buy a raspberry pi but idk what I would do with it 2015-02-09 05:57:13 XgF: if someone donates the hardware, I see no reason why we wouldn't 2015-02-09 09:19:20 Why not an Odroid C1? It has nicer specs like faster CPU and Gigabit ethernet… 2015-02-09 10:55:24 RPI2 have new cpu, but old GPU, and ethernet is still over usb chip so... i dont see much need to upgrade my rpi1. 2015-02-09 11:01:31 true 2015-02-09 11:01:41 crow_, i agree; ethernet over usb is a pain 2015-02-09 11:01:58 and i mostly do only simple omxplayer stuff, so having faster cpu does not mean much for me 2015-02-09 11:10:38 fabled it is same here :/ 2015-02-09 11:11:05 thus Odroid C1 or something like that which have normal ethernet chip would be better, but price is then problem :) 2015-02-09 11:14:04 yeah 2015-02-09 11:14:16 i have wandboards which are nice. but they are quite a bit more expensive too. 2015-02-09 11:28:07 i got my rpi2 2015-02-09 11:28:15 but having power issues 2015-02-09 11:30:16 clandmeter need more Amper? 2015-02-09 11:31:19 i dont know what the problem is 2015-02-09 11:31:27 it seems stable for most 2015-02-09 12:02:37 clandmeter, taking photoes with xenon flash? 2015-02-09 12:02:38 ;) 2015-02-09 12:17:15 :) 2015-02-09 12:18:05 i think i found a proper powersupply 2015-02-09 12:18:08 seems stable now 2015-02-09 14:13:41 i packaged up dalias' compatibility fix for emacs to work with musl, however the thing only works if /proc/sys/kernel/randomize_va_space is 0, and abuild is not running as root. is this a problem? 2015-02-09 14:14:00 without grsec emacs builds fine 2015-02-09 14:14:30 but if someone wants to build emacs on a grsec box, that setting needs to be temporarily disabled. 2015-02-09 14:15:31 tA2p0mM5M4HU, our official builders run grsec kernel 2015-02-09 14:15:47 randomize_va_space can be set on file basis too 2015-02-09 14:15:52 with paxctl 2015-02-09 14:16:11 though, that probably needs makefile patching 2015-02-09 14:18:44 i'll try that 2015-02-09 14:19:39 tA2p0mM5M4HU, paxctl -r 2015-02-09 14:19:45 for the filein question 2015-02-09 14:20:04 you probably need -c or -C too to create the pax flags elf header 2015-02-09 14:22:15 okithx! 2015-02-09 14:26:51 ok, build is running on grsec, should fail in a few minutes then i try the paxctl thing if it works. 2015-02-09 14:35:19 huh, the makefile even contains a PAXCTL_if_present variable \o/ 2015-02-09 14:36:28 yep. seems to work. stupid me not thinking myself about this solution. 2015-02-09 14:37:04 algitbot always happy 2015-02-09 14:38:09 ok. 2nd question. this whole thing only works with the master branch, this doesn't work with the latest release 24.4. so i guess we wait until there is a release for emacs, or is it ok to download cgit .tgz's? 2015-02-09 14:44:29 tA2p0mM5M4HU, we have some git snapshots in testing/ 2015-02-09 14:44:45 i believe we are ok to take them in testing at least 2015-02-09 14:45:02 we have few of them even in main, iirc 2015-02-09 14:45:51 ok, running a final test compile on the x11-less emacs. 2015-02-09 14:46:26 wasn't able to test the bells-and-whistles-x11 version of emacs, due to lack of x11 hosts. 2015-02-09 14:52:10 regardless, sounds very nice. 2015-02-09 16:01:24 here is the light package: https://www.ctrlc.hu/~stef/emacs-nox11.txz 2015-02-09 16:01:45 apkbuild+patches 2015-02-09 16:21:57 tA2p0mM5M4HU, the existing paxctl stuff is not enough? 2015-02-09 16:22:07 and seems the gitweb thingy does not produce identical .tar.gz, so checksumming fais 2015-02-09 16:22:09 fails* 2015-02-09 16:39:06 no they do not set temacs to -r unfortunately. only its product. 2015-02-09 16:39:26 here is the x11 version https://www.ctrlc.hu/~stef/emacs-x11.txz 2015-02-09 16:56:40 i don't quite understand why the gitweb thingies never produce the same archive, yet it is based on the exact same version and content. 2015-02-09 17:00:54 tA2p0mM5M4HU: https://github.com/jomat/aports/blob/master/testing/charybdis-git/APKBUILD seems to work with checksums 2015-02-09 17:05:51 jomat: that looks like a github link, i think the sw that serves emacs, is cgit. that might explain the difference? 2015-02-09 17:06:24 Oh I see... cgit is different of course 2015-02-09 17:32:26 i'll be mostly distracted/afk this week, i hope the apks work out somehow 2015-02-09 19:27:10 clandmeter what is proper psu? 5V and how much A? 2015-02-10 09:41:02 crow: one that can deliver enough voltage/power 2015-02-10 09:41:15 im now using a 5V/2A 2015-02-10 09:48:17 tA2p0mM5M4HU, emacs-x11 does not build here 2015-02-10 09:48:35 various errors from alsa, and other sound stuff 2015-02-10 09:49:09 perhaps my build laptop has additional -dev packages 2015-02-10 09:51:02 ah 2015-02-10 09:51:04 missing linux-headers 2015-02-10 10:54:26 bah 2015-02-10 11:29:51 hmmm. note to self, clean up build vm. sorry fabled /o\ if we ever meet i get you a beer. 2015-02-10 11:30:15 tA2p0mM5M4HU, np. i messed up too with autoconf, and x86-build 2015-02-10 11:30:21 dunno what is broke in x86 2015-02-10 11:30:35 so it's x86-64 only for the time being 2015-02-10 11:30:49 fyi, i will be away from tomorrow until Feb 22nd 2015-02-10 11:30:56 ncopa is away too 2015-02-10 11:31:11 so we will have very limited alpine development for next 1.5 weeks 2015-02-10 11:35:08 oki, i'm away for the rest of the week. 2015-02-10 11:36:14 enjoy your afa time. 2015-02-10 11:37:26 clandmeter: last men standing 2015-02-10 11:38:15 :) 2015-02-10 11:38:29 last man idle 2015-02-10 11:41:47 fabled: x86 is broken atm? 2015-02-10 11:42:27 why? 2015-02-10 11:42:35 only for emacs 2015-02-10 11:52:09 conversation was not clear for me if it was just that build or in general. 2015-02-10 12:02:09 clandmeter, see the commit 2015-02-10 12:02:15 http://git.alpinelinux.org/cgit/aports/commit/?id=2140f80c078bb79fe30c9c59a4edfebfdba15490 2015-02-10 12:02:52 thx got it. 2015-02-10 12:18:32 if i just change the Maintainer of a package, do I need to change the pkgrel? 2015-02-10 12:25:40 nope 2015-02-10 12:25:58 ¬¬" hehehe 2015-02-10 12:26:02 "Always increment pkgrel when making updates to an aport;" 2015-02-10 12:26:28 reading this i just did the pkgrel change 2015-02-10 12:26:29 =/ 2015-02-10 12:26:56 then, that is all that I need to do: http://sprunge.us/KPNE 2015-02-10 12:26:56 ? 2015-02-10 12:27:42 imho yes 2015-02-10 12:28:41 who can apply this patch? I'd like to move vino to main/ 2015-02-10 12:30:55 fabled: have you had a chance to think how to go about integrating fortify into alpine? 2015-02-10 12:36:22 alacerda: why not move it to main in a single commit? 2015-02-10 12:37:56 is it possible? All that i need to do is move it here and send a patch? 2015-02-10 12:38:28 i can also move it if you like. but you can also do it yourself. whatever you prefer :) 2015-02-10 12:38:49 Go ahead. I git pull here 2015-02-10 12:38:56 :) 2015-02-10 12:39:07 :) 2015-02-10 12:39:16 too late, guys 2015-02-10 12:39:18 moving it is faster then typing a msg here 2015-02-10 12:39:27 and i just sent xrdp with the same porpuse: moving to main 2015-02-10 12:39:33 if you can do the same 2015-02-10 12:39:48 i can only do sprunge 2015-02-10 12:40:01 http://sprunge.us/YHhO 2015-02-10 12:40:01 this 2015-02-10 12:40:18 yes 2015-02-10 12:40:27 barthalion: ^ :) 2015-02-10 12:40:52 this one is the XRDP patch 2015-02-10 12:41:02 i have no email config on my builder 2015-02-10 12:41:18 sprunge is so much easier :) 2015-02-10 12:41:21 great! 2015-02-10 12:41:27 np 2015-02-10 12:41:42 Now I am going to write a wiki page about xrdp and vino on AL 2015-02-10 12:41:43 ;) 2015-02-10 13:10:00 stateless, sorry, been busy with other stuff. but it's in my todo list. 2015-02-10 13:10:20 did dalias have any additional feedback on it? 2015-02-10 13:10:33 i wonder how to integrate it to toolchain properly 2015-02-10 13:10:49 it would need to be in some default system include path or similar 2015-02-10 13:10:58 it'd have to come from musl, or gcc 2015-02-10 13:11:13 or at least have some glue in gcc to automatically pick up the path if present 2015-02-10 13:12:25 stateless, and i'm going away for vacation like ncopa. though, fortify is something i'd like to see merged before alpine 3.2 branching 2015-02-10 13:12:54 though, i'd like to have xattr support in apk-tools even before that 2015-02-10 13:12:58 so i'll probably work on that first 2015-02-10 14:13:39 i'm getting WARNING: The repository tag for world dependency 'sj-keys@st.ilet.to' does not exist 2015-02-10 14:13:58 i have a second line in my repo with a private repository 2015-02-10 14:14:32 what's the best way to clean that up? 2015-02-10 14:17:50 looks like my reponame @st.ilet.to is invalid 2015-02-10 14:36:22 fabled: cool! dalias did not have any further feedback (well he did mention a few minor things but nothing to do with the overall structure of fortify) 2015-02-10 14:36:47 stateless, if you or dalias has ideas how to integrate, i'm all for it 2015-02-10 14:37:02 perhaps start asking about that on the mailing list? 2015-02-10 14:37:27 i think simples is to add some #include dir in gcc specs that gets preference over /usr/include 2015-02-10 14:37:28 I have not thought about it, is it easy to simply prepend it to the default include path? 2015-02-10 14:37:33 yes 2015-02-10 14:38:06 i'm also wondering if it should be a) separate package just shipping the headers, b) part of gcc, c) part of musl package 2015-02-10 14:38:15 probably a or c 2015-02-10 14:38:21 and just patch gcc for the include 2015-02-10 14:38:34 I'd be inclined to say a 2015-02-10 14:38:39 metoo 2015-02-10 14:39:07 fortify-headers or similar, and make build-base depend on it 2015-02-10 14:39:10 do you know what the proposed plan is for the compat _chk_* functions? are they going to be in this glibc compat shared lib or part of libc.so? 2015-02-10 14:39:19 it needs to go in libc.so 2015-02-10 14:39:22 ok 2015-02-10 14:39:27 otherwise it's not possible to provide binary compat. 2015-02-10 14:39:34 but dalias wanted those only for the binary compat. 2015-02-10 14:39:43 and native builds would be using the inline headers 2015-02-10 14:39:58 so no references for _chk_ symbols would be generated in our packages 2015-02-10 14:40:13 so some of the existing fortify code can be used for the chk symbols in musl 2015-02-10 14:40:26 for the runtime checks etc. 2015-02-10 14:41:00 yes 2015-02-10 14:41:07 actually, you should ask dalias about that 2015-02-10 14:41:14 what is his plan to provide the _chk_ symbols 2015-02-10 14:41:41 do you have more info on this glibc compat shared lib? I talked to nsz briefly over lunch the other day, he mentioned that it is supposed to be preloaded and even override some of the musl symbols but there were issues that needed to be tackled first (potentially patching the loader amongst other things) 2015-02-10 14:41:42 and would he want to include the headers in musl 2015-02-10 14:41:46 or should they go separate package 2015-02-10 14:42:05 oh 2015-02-10 14:42:12 they wanted to do the binary compat via preload? 2015-02-10 14:42:23 that's my impression 2015-02-10 14:42:34 but I am not sure if there is any work on that atm - I will talk with dalias 2015-02-10 14:42:42 ok 2015-02-10 14:42:52 if possible, use mailing list. i'll follow that discussion then 2015-02-10 14:42:58 i'll be out from irc for next 1.5 weeks 2015-02-10 14:43:19 or at least follow up the email thread on ml 2015-02-10 14:43:29 so i know what's happening / the result of those discussions 2015-02-10 14:43:45 but if they want it a preload lib 2015-02-10 14:43:56 then it'd definitely be external package 2015-02-10 14:44:04 though it's slightly awkward 2015-02-10 14:44:23 perhaps the musl loader could automatically preload it if it detects _chk_ symbols 2015-02-10 14:44:35 I will discuss on the mailing list yes 2015-02-10 14:44:45 though that might have it's own set of problems too 2015-02-10 14:44:50 ok 2015-02-10 14:44:56 let's see what dalias had in mind 2015-02-10 14:46:11 ok 2015-02-10 14:47:50 bye 2015-02-10 16:26:20 seems ncopa has a boring holiday :) 2015-02-10 16:29:21 yeah 2015-02-10 16:31:05 Hi, clandmeter. Did you hear about the effort to bring latex to alpine linux? #alpine-tex 2015-02-10 16:31:27 I think it would be nice to kind of "officialy" make that known. 2015-02-10 16:36:02 clandmeter, copa's wife is gonna kill him :) 2015-02-10 16:37:31 Maybe he has different priorities :> 2015-02-10 17:16:43 please: http://sprunge.us/dGcj 2015-02-10 17:16:48 it is #3742 2015-02-10 17:16:51 :) 2015-02-10 17:37:57 clandmeter ^^ 2015-02-10 17:41:27 abuild -r returned: "Found executable files with SUID bit set: ...." Ok, should I unset the SUID of this file or what? (SUID set is *aways* too dangerous that AL doesn't allow it) 2015-02-10 17:41:28 ? 2015-02-10 18:56:51 alacerda: no, we do not ban use of suid bit 2015-02-10 18:56:59 other packages use it 2015-02-10 18:58:23 tdtrask, >>> ERROR: duo_unix*: Found executable files with SUID bit set: 2015-02-10 18:58:42 what is weird is that it returns as ERROR 2015-02-10 18:59:33 if you really want it, you can add: options="suid" 2015-02-10 18:59:38 to your APKBUILD 2015-02-10 19:00:28 thx tdtrask 2015-02-10 19:01:34 welcome 2015-02-10 19:03:44 looks like about 30 packages in main and another 10 in testing use suid 2015-02-10 22:00:20 yay 2015-02-10 22:00:25 seemed to easy to be real 2015-02-11 11:09:58 alacerda, missing dep fix in awesome also means it needs to be backported to v3.1 ? 2015-02-11 11:10:55 or just let user figure out the dep 2015-02-11 11:13:24 backport would be better 2015-02-11 11:41:17 #3159: http://sprunge.us/WRPV 2015-02-11 11:50:07 alacerda: http://www.redmine.org/projects/redmine/wiki/RedmineSettings#Referencing-issues-in-commit-messages 2015-02-11 11:50:30 alacerda: example: http://git.alpinelinux.org/cgit/aports/commit/?id=5bcd5f35212833b043f22e839a37619579d9b2aa 2015-02-11 11:51:13 i edited to prev commit msg to include it. 2015-02-11 11:51:26 to/your 2015-02-11 12:26:06 clandmeter, thanks 2015-02-11 16:31:41 hi 2015-02-11 16:32:24 i'm happy to announce dl-6.alpinelinux.org ... ip address 107.181.185.116, serving via http, ftp and rsync and now doing the initial sync. just point your dns there and it's available after the init sync 2015-02-11 16:32:38 location is phoneix, usa 2015-02-11 16:34:14 \o/ 2015-02-11 16:34:14 nice 2015-02-11 16:34:24 grimeton: initsync from nl.a.o? 2015-02-11 16:34:36 yeah 2015-02-11 16:34:37 grimeton, cool o/\o 2015-02-11 16:34:46 what is your connection? 2015-02-11 16:34:57 maybe an initsync from a faster mirror is easier 2015-02-11 16:35:01 eh, it's a big dc ... 2015-02-11 16:35:18 i think there are some 1g mirrors 2015-02-11 16:35:49 the machine itself is connected via 10g 2015-02-11 16:36:03 and the dc, don't know ... 300 or 400g at least 2015-02-11 16:36:35 yes, if possible, try an initsync from a faster mirror, nl is only 100mbit 2015-02-11 16:36:53 ok, any mirror near by i should use? 2015-02-11 16:37:47 sure 2015-02-11 16:37:58 i know we have a fast german one 2015-02-11 16:38:48 i remember http://mirrors.gigenet.com/alpinelinux is minum 1g 2015-02-11 16:38:57 do they serve via rsync? 2015-02-11 16:39:41 dl-5 .. my other mirror ... has a bad connection to the dc in phoenix, so don't come up with that one ;) 2015-02-11 16:40:15 if i'd knew the location of the other mirrors ... 2015-02-11 16:40:23 rsync://mirrors.gigenet.com/... 2015-02-11 16:40:58 ok 2015-02-11 16:42:41 if it doesnt work, just use nl 2015-02-11 16:42:53 end of day anyway here :) 2015-02-11 16:42:58 working day that is. 2015-02-11 16:44:25 hm yeah, that one doesn't work ... 2015-02-11 16:45:16 would be nice to have an alternative rsync mirror 2015-02-11 16:45:28 dl-5 and dl-6 both allow rsync 2015-02-11 16:45:35 and anonymous ftp 2015-02-11 16:46:12 i should upgrade our connection. 2015-02-11 16:46:32 seems its technically possible now with new infra from provider. 2015-02-11 16:46:54 100mbit is so 2001 2015-02-11 16:46:57 or we should use something like google drive or something similar for syncing and let them handle the global distribution problem :D 2015-02-11 16:47:15 akamai ftw 2015-02-11 16:48:12 yeah but they want $$, don't they? 2015-02-11 16:48:21 we should have something like master -> slaves -> mirrors, or however the nameing 2015-02-11 16:48:53 yeah, with only 20 mirros at hand ... 2015-02-11 16:49:11 going home. thx for the new mirror. dont forget to put it on the ML 2015-02-11 16:49:18 ehm yeah 2015-02-11 16:49:24 hf 2015-02-11 16:49:32 do you take care of the dns stuff? 2015-02-11 16:49:35 ncopa can add it to the mirrors list 2015-02-11 16:49:49 jeff does that normally 2015-02-11 16:49:53 but dont see him that often. 2015-02-11 16:50:09 i put it on the ml and then we wait 2015-02-11 19:38:16 new mirror? nice :D thank you grimeton 2015-02-12 09:57:49 ERROR: unsatisfiable constraints: 2015-02-12 09:57:51 graphite2-dev (missing): 2015-02-12 09:57:53 required by: world[graphite2-dev] 2015-02-12 09:58:04 since when does the order matter? 2015-02-12 09:58:12 oh its testing 2015-02-12 09:58:15 duh 2015-02-12 09:58:15 yeah 2015-02-12 09:58:19 solly 2015-02-12 09:58:26 you should revert the change in main, and copy harfbuzz to testing 2015-02-12 09:58:33 with new deps 2015-02-12 09:59:55 I wouldn't call it extensive testing they way you did it 2015-02-12 10:01:02 excuse me? 2015-02-12 11:45:40 https://github.com/jomat/aports/blob/master/testing/irssi-otr-git/APKBUILD ... do we want git builds in our aports tree? 2015-02-12 11:49:06 i guess if they produce the same hash 2015-02-12 11:49:44 you coudl also patch 1.0 2015-02-12 11:50:05 cgit uses gzip with timestamps. so the hash never matches 2015-02-12 11:50:54 looks like 1.0 is 10 commits behind 2015-02-12 11:51:06 clandmeter: 1.0? Is there a irssi-otr 1.0 in alpine? 2015-02-12 11:51:07 but github might do not mess up their output 2015-02-12 11:52:14 jomat: i dont know 2015-02-12 11:52:25 but github has 1.0? 2015-02-12 11:53:35 Ah! I only knew ftp://download.tuxfamily.org/irssiotr/ where the latest version is 0.3 from 2009 :-) 2015-02-12 11:54:27 Indeed, there is an 1 yr old 1.0.0 release on Github 2015-02-12 11:58:43 I did the same with charybdis, as the release tarball which is packaged to alpine is years old and misses tons of bugfixes: https://github.com/jomat/aports/blob/master/testing/charybdis-git/APKBUILD 2015-02-12 14:55:47 texlive has pkgdesc="A complete TeX distribution" 2015-02-12 14:56:03 and git comment as "This is a minimal src build. Dont expect it to work at all." 2015-02-12 14:56:05 ;) 2015-02-12 15:12:03 vkrishn: :) 2015-02-12 15:12:17 its in testing for a reason ;-) 2015-02-12 15:12:53 :) 2015-02-12 17:24:51 Hey. I'm working on kamailio APKBUILD for 4.2.3. It will need to be backported to 3.x.x, but the current APKBUILD uses mariadb-dev instead of mysql 2015-02-12 17:25:00 Thinking of doing this as a simple workaround: 2015-02-12 17:25:01 _mysql=mariadb-dev 2015-02-12 17:25:01 [ -z "$( apk search mariadb-dev)" ] && _mysql=mysql-dev 2015-02-12 17:25:02 makedepends="bison flex expat-dev postgresql-dev pcre-dev $_mysql 2015-02-12 17:25:07 thoughts? 2015-02-12 17:25:18 just use mysql-dev in APKBUILD for 3.x 2015-02-12 17:27:36 mmm. then I need to maintain the APKBUILD for the 3.x branch /and/ the master branch... 2015-02-12 17:27:59 it's one time cost I guess 2015-02-12 17:28:15 assuming you don't change the deps in the future, cherry pick will work just fine 2015-02-12 17:28:17 :) kamailio likes to do lots and lots of releases. :) 2015-02-12 17:29:32 doesn't change what I said :p 2015-02-12 17:31:32 yes, thanks. I can't say the dependencies won't change. http://bugs.alpinelinux.org/issues/3919 wants a redis package 2015-02-12 17:31:42 someone else wants libmemcache backported to 3.1 2015-02-12 17:31:59 but thanks for the input 2015-02-12 21:20:51 I would love to have access to the useradd program, as part of the shadow-utils package. I see that it should be available for busybox: http://lists.busybox.net/pipermail/buildroot/2015-January/116899.html 2015-02-12 21:21:25 Is it possible to build alpine with shadow-utils or have it available as a package? 2015-02-12 21:56:14 wat 2015-02-12 21:57:18 gcc bug? 2015-02-12 21:57:52 apparently not 2015-02-12 22:00:33 kaboom again 2015-02-12 22:00:52 and fabled is not here :c) 2015-02-12 22:11:08 :) 2015-02-12 22:11:15 btw. thx for the push 2015-02-12 22:16:56 np 2015-02-13 20:32:23 Hi again. Not sure if this is the right place to ask: Is it possible to build alpine with shadow-utils or have it available as a package? 2015-02-13 20:33:46 yes it is in testing 2015-02-13 20:34:21 so unprivileged LXC containers are possible - just manually create /etc/subuid /etc/subgid 2015-02-13 20:35:11 'apk add shadow' 2015-02-13 21:46:32 BitL0G1c: Nice, how long is the testing usually? Is it just to make a git clone and test for myself, right? 2015-02-13 21:52:18 I think he means add a line to your /etc/apk/repositories file: 2015-02-13 21:52:30 http:///alpine/edge/testing 2015-02-13 21:52:48 then: apk update apk add shadow@edge 2015-02-13 21:52:55 then: apk update apk; add shadow@edge 2015-02-13 21:53:28 mmm 2015-02-13 21:53:37 sorry /etc/apk/repositories should be: 2015-02-13 21:53:45 @http:///alpine/edge/testing 2015-02-13 21:53:52 (the @ at the front of the line is important) 2015-02-13 21:54:06 http://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Repository_pinning 2015-02-13 21:54:30 thanks 2015-02-13 22:23:39 yes do the repo pinning - @atmoz: shadow works ok - I've used it for unprivileged ubuntu server & centos 6 containers. Just "create unprivileged containers as root" (as there is no cgmanager in Alpine) https://linuxcontainers.org/lxc/getting-started/ 2015-02-13 22:25:35 Pinning works like a charm. I'm using docker atm, and it works really well. 2015-02-13 22:30:10 docker is great for running a single app 2015-02-13 22:33:41 Yeah. Intersting things evolving in containerization. Can't wait to see what Rocket will bring to the table 2015-02-16 08:56:31 anyone having a copy of linux-grsec-dev-3.14.27-r2.apk lying around ? 2015-02-16 09:49:58 vkrishn: unless i'm looking in the wrong place, AL 3.0 has linux-grsec-dev-3.14.22-r1.apk and AL 3.1 has linux-grsec-dev-3.14.30-r0.apk 2015-02-16 09:51:30 alpine-3.1.1-x86_64.iso has linux-grsec-3.14.27-r2.apk 2015-02-16 09:53:27 ah, i was looking in the repositories 2015-02-16 09:58:54 only way seems is use v3.1.2 iso 2015-02-16 10:10:32 I rsync'd couple of days before 3.1.2 release 2015-02-16 16:14:24 hmm, what's the prefered method of getting round error.h include failures in musl when building an apk? 2015-02-16 16:15:16 i understand error.h is glibc specific and isn't implemented in musl - could be wrong 2015-02-16 16:18:30 poking around in the repos, i see some people patching source during an abuild to remove error.h includes 2015-02-16 16:18:50 is that acceptable? 2015-02-16 17:03:02 do I need to do anything specific to get KMS to kick in at boot-time? 2015-02-16 17:03:13 on 3.1.2 2015-02-16 17:03:17 using i915 2015-02-16 17:03:31 I am using the vanilla kernel from the repo 2015-02-16 17:04:38 exiting from X gives me a black screen and I can blindly restart it, I think with KMS enabled it should be a sufficient workaround for this issue 2015-02-16 17:06:40 stateless, add fbcon to /etc/modules and reboot 2015-02-16 17:07:00 or modprobe fbcon to test it now without rebooting 2015-02-16 17:07:09 I think that will fix it 2015-02-16 17:10:56 iie: I will test that at home, it will probably work 2015-02-16 17:11:08 good, let me know 2015-02-16 17:11:19 you have an intel video card, right? 2015-02-16 17:12:36 yes, maybe just modprobe i915? 2015-02-16 17:13:18 that could be enough, I don't know 2015-02-16 17:13:27 will try that too :) 2015-02-16 17:13:35 I remember doing this before so I think this should fix it 2015-02-16 17:45:05 iie: works fine :) 2015-02-16 17:47:58 just modprobe i915? 2015-02-16 17:48:18 modprobe i915 gave me a black screen, but modprobe i915 after modprobe fbcon worked 2015-02-16 17:48:23 nice 2015-02-16 17:48:49 now exiting X is fine too (no black screen) 2015-02-16 17:49:04 you should put that on /etc/modules so it persists after reboot 2015-02-16 17:49:20 yeah did that already, rebooted as well 2015-02-16 17:49:22 all good :) 2015-02-16 17:49:26 thanks 2015-02-16 17:49:44 no problem, I had the same issue yesterday after upgrading alpine 2015-02-17 13:05:20 please git pull git://git.alpinelinux.org/fab/aports 2015-02-17 13:05:35 thanks 2015-02-17 13:14:35 Hi all! May be i'm not the first person with this idea, but what do you think about "paxflags" variable in APKBUILD? 2015-02-17 13:15:24 how do they apply to individual executables? 2015-02-17 13:16:21 if you have /bin/a and /bin/b in one pkg of which only /bin/b needs -m, how do you handle that 2015-02-17 13:16:59 Like in md5sums 2015-02-17 13:25:31 why would a simple paxctl -mC /bin/b in the package() be not enough? 2015-02-17 13:26:50 this is worth revisiting if ncopa ever starts rewriting abuild 2015-02-17 13:29:12 paxctl -mC /bin/b in the package() is good too, this is just idea to make buildfile simpler (which means more maintainers IMHO) 2015-02-17 13:40:06 Anyway, thank you guys! Alpine is amazing! 2015-02-17 17:17:28 uggedal, was there a solution to http://www.irclogger.com/.alpine-devel/2014-07-28#1406533329 ? 2015-02-17 17:17:53 or even a temporary one ;) 2015-02-17 20:34:23 hello, i installed alpine to chroot and every time I get back into the chroot apk doesnt work anymore because of "ERROR: Unable to lock database: No such file or directory" 2015-02-17 20:34:54 algitbot: whatis apk database 2015-02-17 20:41:06 has you tries straceing it? 2015-02-17 20:59:20 nope, deleting and rebuilding with strace atm 2015-02-17 21:49:04 faccessat(3, "var/lib/apk/installed", F_OK) = -1 ENOENT (No such file or directory) 2015-02-17 21:53:47 that helped :) I have broken links in /var 2015-02-17 21:55:39 mkdir /run/lock fixes it 2015-02-18 09:03:18 barthalion: have you used yum on alpine? 2015-02-18 09:08:37 I don't think so 2015-02-18 09:08:47 have I packaged it? 2015-02-18 09:09:01 yes 2015-02-18 09:09:11 trying to create a centos container 2015-02-18 09:09:46 some user requested it 2015-02-18 10:09:03 barthalion: yum is ok, it was grsec kicking in. 2015-02-18 14:35:49 I got texlive to compile on alpine, finally 2015-02-18 14:35:53 ! 2015-02-18 14:36:02 i'm still strugling with ksh 2015-02-18 14:36:05 it just needed some small modifications because of musl 2015-02-18 14:36:10 and some fiddling around 2015-02-18 14:36:19 for anyone interested, here's a dirty guide 2015-02-18 14:36:22 http://sprunge.us/VEYh 2015-02-18 14:37:14 nice. you could package this up easilyt 2015-02-18 14:37:23 s/t$// 2015-02-18 14:37:40 could I? I'll have a read on the wiki page about pkgbuild 2015-02-18 14:39:54 this is the ref: http://wiki.alpinelinux.org/wiki/APKBUILD_Reference 2015-02-18 14:40:01 and some nice examples: http://wiki.alpinelinux.org/wiki/APKBUILD_examples 2015-02-18 14:40:01 I'm starting vacation too, so ask clandmeter for anything you need 2015-02-18 14:40:08 will be back in a week 2015-02-18 14:40:17 barthalion, came to Brazil 2015-02-18 14:40:17 okay, have a nice one 2015-02-18 14:40:18 :P 2015-02-18 14:40:23 enjoy your holidays! 2015-02-18 14:41:33 thx, and see you 2015-02-18 19:31:13 hmm, i get "2.1.1.1 is not a valid version" when doing `abuild -r` 2015-02-18 19:49:15 why 2.1.1.1? 2015-02-18 19:50:37 nothing suprising that don't work with abuild. 2015-02-18 20:08:21 why? 2015-02-18 20:12:53 what why? post your APKBUILD here 2015-02-18 20:20:37 i wonder because 2.1.1.0 was considered valid 2015-02-18 20:25:32 http://sprunge.us/djAU 2015-02-18 20:30:44 works for me. 2015-02-18 20:31:32 strange. 2015-02-18 20:31:35 now here to 2015-02-18 20:31:38 o 2015-02-18 20:31:41 -o 2015-02-18 20:32:23 thx! and sorry for my confusion 2015-02-19 11:17:29 clandmeter [vdr] [ANNOUNCE] VDR version 2.2.0 released - Celebrating 15 years of VDR! http://www.linuxtv.org/pipermail/vdr/2015-February/028675.html 2015-02-19 11:47:27 anyone working with https://bugs.alpinelinux.org/issues/2972 ? 2015-02-19 11:54:53 kwork i think few of main dev are on vication (ncopa barthalion ), but what i remmember fabled also worked on nfs-utils 2015-02-19 11:55:11 crow: nice, but still prefer tvh :) 2015-02-19 15:57:12 rpcbind didnt segfault for me on this version. so if anybody wants to try nfs. 2015-02-19 19:42:26 9/buffer #wht 2015-02-20 10:12:18 hm, I screwed up or buildrepo isn't a part of alpine-sdk? I build xorg-server 1.17.1, but I can't rebuild rest of packages against new version. 2015-02-20 10:15:32 i dont think its part of sdk, atleast not on my system 2015-02-20 10:20:37 scadu: buildrepo is provided by lua-aports now 2015-02-20 10:22:26 clandmeter: good to know, thanks. 2015-02-20 10:30:59 but I think that previously packages were installed from packages directory if their version was higher than these in repos. 2015-02-20 10:31:02 hm… 2015-02-23 07:32:42 bah 2015-02-23 09:29:40 not my day today 2015-02-23 10:40:59 fabled: do you use checkapk for soname changes 2015-02-23 10:42:11 yes, when doing major update 2015-02-23 10:45:31 fabled: can i trust it? i tried to update dbus and got this http://sprunge.us/USJi 2015-02-23 10:45:49 yes, that's ok 2015-02-23 10:46:07 soname is like libdbus-1.so.3 2015-02-23 10:46:08 ok 2015-02-23 10:46:35 then ill just push it 2015-02-23 10:46:41 sure 2015-02-23 10:46:42 thanks 2015-02-23 11:00:12 so Linux 4.0 is on its way 2015-02-23 11:02:28 yes, -rc1 is tagged 2015-02-23 13:44:34 anyone here use jackd? 2015-02-23 14:16:28 rnalrd: i sent in 3 patches around January 31 that weren't applied yet, which explains why py-exifread is missing the changes for 2.0.0 2015-02-23 14:17:08 actually i sent in 4 patches: mdocml, youtube-dl, py-exifread, and py-twitter 2015-02-23 14:17:15 sorry 2015-02-23 14:17:27 let me see i have them somewhere 2015-02-23 14:17:32 if I have* 2015-02-23 14:19:25 pnutzh4x0r, sorry can't find them in my box 2015-02-23 14:19:38 do you mind in resending them? 2015-02-23 14:19:48 no problem 2015-02-23 14:19:53 directly to my email if you wish 2015-02-23 14:19:59 ok 2015-02-23 14:20:03 or ML a you prefer 2015-02-23 14:20:04 tnx 2015-02-23 14:25:17 pnutzh4x0r: found the patches already on the ML 2015-02-23 14:25:58 http://lists.alpinelinux.org/alpine-aports/ - they are all there 2015-02-23 14:26:10 to get them sent to you again as mail (and not HTML) 2015-02-23 14:26:22 pnutzh4x0r, looks like I wasn't subscribed yet to the new ML when you sent those 2015-02-23 14:26:26 ok nvm 2015-02-23 14:26:30 just got your email 2015-02-23 14:26:36 send an email to by sending an email to alpine-aports+get-N@lists.alpinelinux.org 2015-02-23 14:27:04 for in this case its: mailto:alpine-aports+get-128@lists.alpinelinux.org 2015-02-23 14:27:27 ailto:alpine-aports+get-129@lists.alpinelinux.org 2015-02-23 14:27:32 ailto:alpine-aports+get-130@lists.alpinelinux.org 2015-02-23 14:27:35 ailto:alpine-aports+get-131@lists.alpinelinux.org 2015-02-23 14:27:55 the mdocml may need to be reviewed since i modified upstream 2015-02-23 14:27:57 pnutzh4x0r, regarding mdocml I'll let ncopa ACK it since he's the maintainer 2015-02-23 14:28:01 right 2015-02-23 14:30:11 nangel, tnx good to know 2015-02-23 14:30:51 ImportError: No module named setuptools 2015-02-23 14:30:51 >>> ERROR: py-exifread: all failed 2015-02-23 14:31:03 pnutzh4x0r ^^^^ 2015-02-23 14:31:40 just FYI 2015-02-23 14:31:57 i'm adding py-setuptools to makedepends 2015-02-23 14:34:47 rnalrd: hmm, ok thnx 2015-02-23 14:35:08 all done 2015-02-23 14:35:13 tnx and sorry for the oversight 2015-02-23 14:35:35 no problem 2015-02-23 14:40:19 morning 2015-02-23 14:41:10 anyone recently tried to boot a xen dom0 via nfs? with alpine 2.7/xen 4.3 it worked fine, now with alpine 3.1.2 and xen 4.4.1 it's lacking nfs capabilities even tho i have it in my initramfs and modloop 2015-02-23 16:02:48 fabled: does x86_64 have headers installed? 2015-02-23 16:03:45 oh it also broke on 64 2015-02-24 08:25:59 hrmpf 2015-02-24 11:01:01 could someone pls bump task to v2.4.1 2015-02-24 11:01:17 it would build after removing the patch "musl-mktime.patch" 2015-02-24 11:01:39 as this patch is now updated upstream 2015-02-24 11:02:05 also APKBUILD need the line refering to "musl-mktime.patch" removed 2015-02-24 11:05:16 thanks 2015-02-24 15:15:36 I sent up a patch for logtail-v3.sf.net; and suggested the package be called "logtail" 2015-02-24 15:16:03 the complication is that "logtail" was the name of a now defunct psionic.com (now cisco) product 2015-02-24 15:16:39 Debian re-implemented the thing in Perl, and calls it logtail2 (but the .deb is named logtail, and the perl script is called logtail) 2015-02-24 15:17:25 then the sourceforge project took the original logtail added 64bit support, and decided to call the project "logtail-v3" 2015-02-24 15:17:37 but the binary object is still called "logtail" 2015-02-24 15:17:50 so what to call the .apk? 2015-02-24 15:18:19 it would technically be logtail-v3-21.r0.apk 2015-02-24 15:18:34 but that looks ugly, esp since the program is called "logtail", not "logtail-v3" 2015-02-24 15:19:38 at the risk of being technically incorrect, I suggested "logtail-3.21-r0.apk" 2015-02-24 15:19:52 anyone have any major problems with that, let me know... 2015-02-24 23:30:44 hm so xen 4.4.1 has a bug i'm hitting (https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=776237), this is the only one i find for my stack trace 2015-02-25 16:52:57 hi, I noticed with Alpine Linux GNU acct (main/acct) "last" will report "/dev/null/wtmp" not found, because it uses musl _PATH_WTMP definition in musl/include/paths.h 2015-02-25 16:53:01 in the acct build process the path is defined in files.h 2015-02-25 16:59:38 ok nvm, its probably intentional (utmp functions are stubs), sorry :) 2015-02-26 11:47:35 rnalrd, what pkgs is using opensp, openjade ? 2015-02-26 11:54:11 fossil binary size is now double its previous version, http://www.fossil-scm.org/download.html 2015-02-26 11:54:13 ;) 2015-02-26 11:54:26 vkrishn, yes 2015-02-26 11:55:04 but I need openjade just for building another package 2015-02-26 11:55:11 thanks for pkgs, opensp, openjade, but could not find what application use them 2015-02-26 11:56:33 docs for https://wiki.postgresql.org/wiki/BDR_Administration 2015-02-26 11:57:03 but looks like I built them in vain as they throw tons of errors and I eventually decided not to build docs at all 2015-02-26 11:57:14 would it be ok if opensp is build with --enable-http ? 2015-02-26 11:57:23 or, --enable-http --disable-doc-build 2015-02-26 11:57:43 np 2015-02-26 11:58:23 was wanting to have validator in AL, http://validator.w3.org/docs/install.html#install-prereq 2015-02-26 12:14:35 mm getting some other compiling issues now 2015-02-26 12:14:52 i'll do it later 2015-02-26 12:14:55 need to go 2015-02-26 12:18:05 ok 2015-02-27 08:44:24 bah 2015-02-27 08:44:41 why didn't I test build before pushing? 2015-02-27 08:51:47 will AL v3.2 be having ruby 2.2.0 ? 2015-02-27 12:21:40 hi, does alpine support the rasperry pi2 ? 2015-02-27 12:33:36 not yet I guess? I think we don't have a kernel for it 2015-02-27 12:35:39 ok ty 2015-02-27 14:56:33 hello algitbot 2015-02-27 20:17:07 spock is dead :( 2015-02-27 20:17:33 🖖 2015-02-28 18:17:05 hm, any ideas why starting /etc/init.d/rpc.statd will get stuck in wait4()? Is it openrc waiting for the child or? 2015-02-28 18:17:11 starting rpc.statd manually works fine 2015-02-28 18:17:24 in fact rpc.statd starts regardless but openrc seems to get stuck in the call waiting 2015-02-28 18:17:50 this is on -edge, didn't happen about two weeks ago.. hm I will investigate some more (I am totally unfamiliar with openrc)