2014-12-01 08:05:00 kaniini, ncopa: i'm trying to build Rasperry Pi stuff without fbcon 2014-12-01 08:05:12 so i get nice boot splash without any blinking cursors 2014-12-01 08:05:31 seems ncopa added loading fbcon via mdev.conf in May 2011 as response to http://bugs.alpinelinux.org/issues/622 2014-12-01 08:05:58 and kaniini added it in initramfs-init just a bit earlier in March 2011 2014-12-01 08:06:29 now i'm wondering what would be the easy way to allow *not* loading fbcon even though fb0 exists 2014-12-01 08:07:12 the bootsplash stuff i just added works, but for a moment one sees blinking cursors (and in some cases a glimpse of apk progress bar) 2014-12-01 08:07:32 due to fbcon gets loaded? 2014-12-01 08:07:37 though, i did patch fbsplash to change consoles so that anything else gets hidden 2014-12-01 08:07:40 ncopa, two reasons 2014-12-01 08:08:08 fbcon loaded shows blinking cursor 2014-12-01 08:08:08 it's possible to also hide the cursor with boot parameter, but that's slightly hacky 2014-12-01 08:08:32 apk progress bar you see for a 1/2 second due to fbsplash being started as background with & 2014-12-01 08:09:14 so apk might get a head start and print something before fbsplash has time to change consoles and init fb0 2014-12-01 08:09:44 so i'm thinking to delete the explicit load of fbcon from initramfs-init 2014-12-01 08:09:50 and add -b to modprobe in mdev.conf 2014-12-01 08:09:55 and a way to pass blacklist to init 2014-12-01 08:10:01 would that sound reasonable? 2014-12-01 08:10:13 i think so yes 2014-12-01 08:10:15 that way 2014-12-01 08:10:18 if i need to debug 2014-12-01 08:10:28 i just remove the 'blacklist' bootarg. and i get fullconsole 2014-12-01 08:10:32 the explicit way we currently load fbcon is hackish imho 2014-12-01 08:10:40 ok i'll remove that now 2014-12-01 08:10:54 done 2014-12-01 08:11:33 i'm also adding -b to modprobe fbcon in mdev.conf 2014-12-01 08:12:16 yes 2014-12-01 08:12:51 done 2014-12-01 08:13:10 so now to just figure out how to blacklist fbcon during the early boot stage? 2014-12-01 08:14:08 for i in ${KOPT_blacklist/,/ }; do echo $i >> /etc/modprobe.d/blacklist.conf; done 2014-12-01 08:14:45 for i in ${KOPT_blacklist/,/ }; do echo "blacklist $i" >> /etc/modprobe.d/boot-opt-blacklist.conf; done 2014-12-01 08:15:18 should that be also copied to $sysroot, or i need to blacklist fbcon both in boot args, and overlay? 2014-12-01 08:15:42 im thinking of that.... 2014-12-01 08:16:00 i think we can copy the boot-opt-blacklist.conf 2014-12-01 08:16:34 which is why i think we should call it boot-opt-blacklist.conf and not just 'blacklist.conf' 2014-12-01 08:17:14 i think that makes sense 2014-12-01 08:18:03 i'm testing the blacklist first 2014-12-01 08:18:14 just simple as http://sprunge.us/QBWi 2014-12-01 08:19:00 like you suggested 2014-12-01 08:19:38 add it to the KOPT whitelist too 2014-12-01 08:22:31 oh right 2014-12-01 08:24:53 hum 2014-12-01 08:25:03 ok that worked, but now fbsplash broke 2014-12-01 08:27:59 hummm 2014-12-01 08:28:05 it seems that fbsplash requires fbcon to work :-o 2014-12-01 08:29:18 huh 2014-12-01 08:32:00 fabled: could you please git pull alpine-iso 2014-12-01 08:32:08 and verify it does not break the rpi generation 2014-12-01 08:33:17 also, how do you generate the rpi image? just 'make rpi'? 2014-12-01 08:33:27 ncopa, yes, 'make rpi' like in the commit comment 2014-12-01 08:33:42 well, 'make rpi PROFILE=alpine-rpi' to be exact 2014-12-01 08:33:55 that was my exact question :) 2014-12-01 08:37:27 oh, i have one more addition to mdev.conf 2014-12-01 08:39:25 fabled: coudl you please check if this lets you do 'make rpi' without the PROFILE=alpine-rpi http://sprunge.us/fNJN 2014-12-01 08:40:04 it should also be possible to run: make all-release 2014-12-01 08:41:21 $ make all-release 2014-12-01 08:41:22 make: *** No rule to make target 'previous', needed by 'all-release'. Stop. 2014-12-01 08:42:05 make release tries wrong thing 2014-12-01 08:43:04 'make rpi' seems to work 2014-12-01 09:04:52 aha, i know what's wrong with fbsplash without fbcon 2014-12-01 09:05:00 fbcon allocates larger virtual screen for scrollback 2014-12-01 09:10:25 i have a new git diff to fix alpine-iso: 2014-12-01 09:10:30 ok? 2014-12-01 09:10:31 git checkout Makefile 2014-12-01 09:10:39 http://sprunge.us/CRJZ 2014-12-01 09:11:09 i think it actually will fix a problem with x.y.0 releases for me too 2014-12-01 09:11:24 when the 'previous' file does not exists 2014-12-01 09:11:35 i wonder if i should drop the pkgdiff and xdelta stuff too 2014-12-01 09:11:41 does anyone acutally use that? 2014-12-01 09:12:02 i've used sometimes 2014-12-01 09:12:05 but not often 2014-12-01 09:12:19 xdelta only, that is 2014-12-01 09:12:43 never used pkgdiff; i'm usually good with the announce mail / website news post 2014-12-01 09:12:58 i sued to use it for announcement 2014-12-01 09:12:58 yes, 'make all-release' looks a lot better now 2014-12-01 09:13:03 used* 2014-12-01 09:13:06 but i longer do 2014-12-01 09:13:09 yeah, now it's from git 2014-12-01 09:13:20 i think git shortlog is better 2014-12-01 09:13:26 i'm letting make all-release run to see if it works 2014-12-01 09:13:35 at least it started to do the right thing 2014-12-01 09:13:47 make -j4 all-release 2014-12-01 09:16:08 mmm 2014-12-01 09:16:11 -j4 left out 2014-12-01 09:16:25 i'm building same time, and it's slow arm 2014-12-01 09:16:33 :) 2014-12-01 09:17:35 i wonder if the rpi portion should use 'cp -l' instead of regular cp 2014-12-01 09:18:08 seems all-release worked; i got .tar.gz + sha signatures of it 2014-12-01 09:18:22 nice! 2014-12-01 09:24:30 hum 2014-12-01 09:24:38 apk bug 2014-12-01 09:24:40 i think 2014-12-01 09:24:53 ncdev-edge-x86_64:~/alpine-iso$ apk apk fetch --keys-dir /etc/apk/keys --repos 2014-12-01 09:24:54 itories-file /etc/apk/repositories --output /home/ncopa/alpine-iso/isotmp.alpine 2014-12-01 09:24:54 ncdev-edge-x86_64:~/alpine-iso$ echo $? 2014-12-01 09:24:54 0 2014-12-01 09:25:08 wrong paste 2014-12-01 09:25:56 apk fetch --keys-dir /etc/apk/keys --repositories-file /etc/apk/repositories --output /home/ncopa/alpine-iso/isotmp.alpine/isofs/apks/x86_64 --recursive aslkdjf lakjh lkjh; echo $? 2014-12-01 09:26:01 0 2014-12-01 09:26:16 apk fetch invalid 2014-12-01 09:26:52 $ apk fetch invalid; echo $? 2014-12-01 09:26:52 0 2014-12-01 09:27:24 that basically means, if packagelist is broken alpine-iso will not catch it 2014-12-01 09:32:15 mm 2014-12-01 09:32:31 sounds likee that 2014-12-01 09:42:51 ncopa: hi, could you please submit a patch for backuppc -> http://sprunge.us/IhhE ? 2014-12-01 09:50:04 k0r10n, i'll take care of it 2014-12-01 09:50:16 rnalrd: thanks! 2014-12-01 09:51:30 i just gave it a quick look and i didnt find anything that immediately caught my attention 2014-12-01 09:51:46 + perl-libwww perl-file-rsync samba-client" 2014-12-01 09:52:07 you cannot run backuppc at all without samba-client install? 2014-12-01 09:52:17 it will fail to run without samba-client in all configurations? 2014-12-01 09:52:20 you can 2014-12-01 09:52:22 ncopa: initd failed to start without it 2014-12-01 09:52:34 it's needed for backup windows machine 2014-12-01 09:52:51 but it you don't have win machines it's useless to have it 2014-12-01 09:53:03 i really don't like backup pc packaging 2014-12-01 09:53:08 so defautl config expects windows machines? 2014-12-01 09:53:18 it looks mess yes :) 2014-12-01 09:53:24 (which is why i prefer not touch it ;) 2014-12-01 09:53:25 i'd really pass maintainership to someone else : 2014-12-01 09:53:32 :-D 2014-12-01 09:53:38 since i even don't use it 2014-12-01 09:54:26 testing build after patching here 2014-12-01 09:56:13 i also noticed there is undefined variable in /etc/BackupPC/config.pl after installing - $Conf{BackupPCUser} 2014-12-01 09:56:23 that's why i added it to patch 2014-12-01 09:57:15 it seems fbcon is required to make bootsplash work :/ 2014-12-01 09:57:30 so i need to figure out how to fix it nicely other than disabling modprobe of fbcon 2014-12-01 11:37:51 hmm... nice git v2.2.0, lots of changes 2014-12-01 11:42:17 ScrumpyJack, any progress on http://lists.alpinelinux.org/alpine-devel/4314.html ? 2014-12-01 12:49:56 hmm 2014-12-01 12:50:09 seems fbplash not working on rpi without fbcon is bug in rpi fb kernel driver 2014-12-01 12:51:42 filed a bug https://github.com/raspberrypi/linux/issues/731 2014-12-01 13:12:44 could someone remind me why abuild won't find the -doc subpackage >>> ERROR: frotz-doc*: Missing /home/scrumpyjack/frotz/pkg/frotz-doc ? 2014-12-01 13:12:57 ncopa, morning! Do you mind to move nikto and hping3 from testing to main? I've tested both of them last friday. 2014-12-01 13:13:38 i've seen this before but i can't remember. it's to do with putting the man page into a package, obviously, but it should "just work" :) 2014-12-01 13:14:55 for i in doc man info html sgml licenses gtk-doc ri help; do 2014-12-01 13:15:30 i guess "$pkgdir/usr/share/$i" need to exist 2014-12-01 13:17:36 install -m755 -d "${pkgdir}/usr/share/man/" fixed it 2014-12-01 13:17:54 sorry for the noise, just a brain fart 2014-12-01 13:33:23 ncopa: can you please publish acf-mysql on git.a.o? 2014-12-01 13:35:21 tdtrask: moved 2014-12-01 13:36:21 and the hook created? 2014-12-01 13:39:08 can you do it? 2014-12-01 13:41:05 ok i did it 2014-12-01 13:41:25 git:/var/git/acf-mysql.git# cp -a ../acf-core.git/hooks/post-receive hooks/ 2014-12-01 13:45:15 ncopa: I can't set it up properly without root permissions 2014-12-01 13:45:25 the hook? 2014-12-01 13:45:33 yeah 2014-12-01 13:45:52 i think you only need write acess to acf-mysql.git/hooks dir 2014-12-01 13:45:59 I had tried it once before (probably a year ago) 2014-12-01 13:46:12 and it didn't work unless the hook itself was owned by root 2014-12-01 13:46:54 maybe I'm wrong, as it doesn't have root ownership now 2014-12-01 13:47:22 we should move to gitolite 2014-12-01 13:47:41 but not today... 2014-12-01 13:50:14 hook seems to work :) 2014-12-01 13:50:19 thanks ncopa 2014-12-01 13:54:47 fabled: ok to tag rc4 now? 2014-12-01 13:54:57 or do you have any more rpi changes in the queue? 2014-12-01 13:55:23 i have a kernel patch pending for rpi frame buffer 2014-12-01 13:55:25 but 2014-12-01 13:55:30 it can wait for _final 2014-12-01 13:55:36 i think i'm ok for _rc4 2014-12-01 13:56:53 i'll finish up other stuff tomorrow 2014-12-01 13:56:55 or late today 2014-12-01 13:57:07 but it's not_rc4 blocker 2014-12-01 13:57:17 mmm 2014-12-01 13:57:20 though 2014-12-01 13:57:26 the blacklist feature should go out 2014-12-01 13:57:28 ncopa, ^^^ 2014-12-01 13:57:58 http://sprunge.us/JHYP 2014-12-01 13:58:47 ok to push that? 2014-12-01 13:59:11 pushed 2014-12-01 13:59:17 you can revert if needed 2014-12-01 14:02:47 did you test it? :) 2014-12-01 14:33:19 ncopa: ncopa, morning! Do you mind to move nikto and hping3 from testing to main? I've tested both of them last friday. 2014-12-01 14:33:33 ncopa: do you want me to do that? 2014-12-01 14:52:18 sure, thanks 2014-12-01 14:55:44 alacerda: ^^^ 2014-12-01 15:01:09 gtk-doc fails to build on armhf 2014-12-01 15:01:19 it fails to find docbook-xml something 2014-12-01 15:04:33 tdtrask, thanks 2014-12-01 15:14:03 ERROR: hping3: Failed to build 2014-12-01 15:14:04 whatta 2014-12-01 15:17:02 is hping3 even maintained upstream? 2014-12-01 15:18:30 finally 2014-12-01 15:19:27 ncopa: have you had time to think about the text i sent for the about page? or was the text a bit ambitious 2014-12-01 15:20:07 only thing i have room for in my small mind atm is: "get v3.1 out the door" 2014-12-01 15:20:16 :) 2014-12-01 15:20:18 its time too shoot the engineers and just ship it 2014-12-01 15:20:49 anything not building? I am free this evening 2014-12-01 15:21:07 hping3 is broke somehow 2014-12-01 15:21:09 ScrumpyJack, gammu ? 2014-12-01 15:21:14 i moved it to unmaintained 2014-12-01 15:21:33 vkrishn: what's missing for the gammu package? 2014-12-01 15:21:39 ncopa: anything else? 2014-12-01 15:21:48 I don't know is it in testing ? 2014-12-01 15:21:57 i havent even started look at the current bugs... 2014-12-01 15:22:10 saw the ML patch, but not in git ? 2014-12-01 15:23:46 vkrishn: it's here @st.ilet.to http://st.ilet.to/alpine 2014-12-01 15:23:54 add that to repositories 2014-12-01 15:28:04 ncopa, rpi release did not get uploaded... 2014-12-01 15:28:30 ScrumpyJack, I can try check the build if patch gets into git 2014-12-01 15:28:38 maybe suggest more changes 2014-12-01 15:28:55 ncopa, seems it got built 2014-12-01 15:28:58 just not uploaded 2014-12-01 15:29:24 it's on builder ~buildozer/alpine-iso 2014-12-01 15:29:35 maybe it was generated to wrong dir? 2014-12-01 15:30:08 need to run 2014-12-01 15:32:26 ncopa, weird on friday i installed hping3 and used here w/o any problem 2014-12-01 15:34:59 its upstream source url 2014-12-01 15:35:12 we have probably had it in our cache since upstream existed 2014-12-01 15:55:58 "release 3.1.0_rc4 is out" congrats! \o/ 2014-12-01 16:03:20 ok, this is what I was thinking, 2014-12-01 16:03:21 3.0.x - musl integration release 2014-12-01 16:03:22 3.1.x - musl related cleanup package cleanups 2014-12-01 16:03:22 3.2.x - next best AL release 2014-12-01 16:04:59 I think, in a rolling release like AL there needs be occassional cleanup release (very 4 cycles or so) 2014-12-01 17:47:51 ncopa ping 2014-12-02 06:55:28 jarruda: pong 2014-12-02 07:33:24 fabled: arm 3.1 builder ahs problems with gtk-doc: 2014-12-02 07:33:28 configure:12126: result: /usr/bin/xmlcatalog 2014-12-02 07:33:28 configure:12145: checking for DocBook XML DTD V4.3 in XML catalog 2014-12-02 07:33:28 configure:12148: $XMLCATALOG --noout "$XML_CATALOG_FILE" "-//OASIS//DTD DocBook 2014-12-02 07:33:28 XML V4.3//EN" >&2 2014-12-02 07:33:28 No entry for PUBLIC -//OASIS//DTD DocBook XML V4.3//EN 2014-12-02 07:33:29 configure:12151: $? = 4 2014-12-02 07:33:31 configure:12157: result: not found 2014-12-02 07:33:33 configure:12159: error: could not find DocBook XML DTD V4.3 in XML catalog 2014-12-02 07:34:41 hmm 2014-12-02 07:36:00 ncopa, talking about arm and the release. i saw files names *-3.1.0_rc4-armhf.iso.sha1 2014-12-02 07:36:09 so it's still trying to create the diff checksums 2014-12-02 07:36:44 seems that's doing of "mklatest.sh" 2014-12-02 07:36:47 did you git pull? 2014-12-02 07:36:52 yes 2014-12-02 07:36:53 oh 2014-12-02 07:37:00 i know why 2014-12-02 07:37:03 seems both, makefile and mklatest.sh try to create checksums 2014-12-02 07:37:28 possibly also something else 2014-12-02 07:37:42 its mklatest.sh 2014-12-02 07:37:45 i forgot fix it 2014-12-02 07:37:58 i fixed 'abuild -A' already on it 2014-12-02 07:38:02 so git pull first 2014-12-02 07:39:32 http://sprunge.us/DNJG 2014-12-02 07:39:55 can you try that, and do mklatest.sh? 2014-12-02 07:48:57 ncopa, it ran, rc=0, latest-releases.yaml is empty 2014-12-02 07:49:26 (well it contains a single line: '---') 2014-12-02 07:49:48 do you have anything in the file 'current'? 2014-12-02 07:50:17 http://sprunge.us/KVET 2014-12-02 07:50:27 $ cat current 2014-12-02 07:50:27 3.1.0_rc4 2014-12-02 07:50:33 $ cat previous 2014-12-02 07:50:33 3.0.0 2014-12-02 07:51:02 ah 2014-12-02 07:51:10 the file suffix is .rpi.tar.gz 2014-12-02 07:51:13 not just .tar.gz 2014-12-02 07:51:27 right 2014-12-02 07:51:29 i currently look for $current-$arch.tar.gz 2014-12-02 07:51:35 hm 2014-12-02 07:51:51 changing that 2014-12-02 07:51:52 http://sprunge.us/GHcI 2014-12-02 07:52:15 looks good 2014-12-02 07:52:18 flavor is not set 2014-12-02 07:53:56 talking on fixing bugs... i still get "GDBus.Error:org.freedesktop.PolicyKit1.Error.Failed: Cannot determine user of subject" error popup on everytime i login on my laptop 2014-12-02 07:54:23 try this: http://sprunge.us/YfAj 2014-12-02 07:55:33 looks good to me 2014-12-02 07:55:37 fabled: do you have system dbus service running? 2014-12-02 07:55:43 did the patch work? 2014-12-02 07:55:54 ncopa, yes, looks like mklatest.sh works now 2014-12-02 07:55:59 yes, system dbus runs: 2014-12-02 07:55:59 1552 ? Ss 0:00 /usr/bin/dbus-daemon --system 2014-12-02 07:56:33 pushed new alpine-iso 2014-12-02 07:56:41 we need make rc5 then :-/ 2014-12-02 07:57:19 mmm 2014-12-02 07:57:28 yeah, might need to do that 2014-12-02 07:57:33 lemme test few things first 2014-12-02 07:57:43 seems the rpi image is missing omxplayer 2014-12-02 07:57:58 could you please doublecheck the mklatest i pushed to alpine-iso? 2014-12-02 07:58:20 yeah, looks good still 2014-12-02 07:58:28 missing package is probably due to apk fetch always returning success 2014-12-02 08:00:48 ncopa, hmm 2014-12-02 08:00:54 that seems to be same issue as with "apk del" 2014-12-02 08:00:58 the wildcard support 2014-12-02 08:01:19 made it so that giving a name matching nothing is not an error 2014-12-02 08:01:41 should probably check that if it does not contain wildcards (*, ?), it needs to match, or error is reported 2014-12-02 08:01:53 yes 2014-12-02 08:02:05 if (has_wildcard(pkg) .... 2014-12-02 08:02:05 i think we have bug about the same issue in apk del 2014-12-02 08:02:33 i think its too late for v3.1 2014-12-02 08:02:47 changing that behavior can have unexcepcted consequences 2014-12-02 08:02:52 that's true also 2014-12-02 08:03:07 but it kinda makes builder defunct: it can build broken images 2014-12-02 08:03:08 morning 2014-12-02 08:12:16 ncopa, hmm, it'd be slightly non-trivial change 2014-12-02 08:12:22 to fix the fetch/del issue 2014-12-02 08:12:30 fetch should report error if download fails,though 2014-12-02 08:13:22 ncopa, i have also one mkinit fix pending 2014-12-02 08:13:27 ok? 2014-12-02 08:13:31 on splash 2014-12-02 08:13:35 i need to test it first 2014-12-02 08:20:37 fabled: btw 2014-12-02 08:20:44 alpine-rpi-3.1.0_rc4-armhf.rpi.tar.gz 2014-12-02 08:20:55 do we need the rpi twice? 2014-12-02 08:21:27 mmm... probably no 2014-12-02 08:22:11 alpine-wanboard-$version-armhf.tar.gz 2014-12-02 08:22:21 alpine-hummingboard-$version-armhf.tar.gz 2014-12-02 08:22:36 i'd prefer alpine-armv6-$version-armhf.tar.gz 2014-12-02 08:22:46 or so 2014-12-02 08:22:57 it's just the bootloader image, and dtb that will be different on those 2014-12-02 08:23:02 the are small files 2014-12-02 08:23:04 but will be many 2014-12-02 08:23:06 hmm, so the checksum doesn't match the file in frotz APKBUILD? 2014-12-02 08:23:21 but they need idfferent bootloader, so we need idfferent images? 2014-12-02 08:23:26 we can think of that later 2014-12-02 08:23:42 but alpine-rpi-$version-armhf.tar.gz works for now 2014-12-02 08:24:12 then its: alpine-$releaseflavor-$version-$arch.tar.gz 2014-12-02 08:24:23 ScrumpyJack: correct 2014-12-02 08:24:47 what's the best way to send a patch or a patch that hasn't been applied? 2014-12-02 08:24:52 s/or/of 2014-12-02 08:25:10 rebase it against current git 2014-12-02 08:25:18 ok 2014-12-02 08:25:33 git send-email --subject-prefix "PATCH v2" 2014-12-02 08:25:38 or so 2014-12-02 08:25:58 ah that frotz patch *has* been applied right? 2014-12-02 08:26:04 no 2014-12-02 08:26:30 many good principles here: http://wiki.qemu.org/Contribute/SubmitAPatch 2014-12-02 08:26:45 how did you tell that the checksum didn't match, during a build? 2014-12-02 08:26:47 it seems we need to write our own "how to use git properly with alpine development" 2014-12-02 08:26:53 yes 2014-12-02 08:27:16 i wanted to verify that it doesnt break the builders 2014-12-02 08:27:18 so i test build 2014-12-02 08:27:22 so you build it to check it works before you apply the patch? that makes sense 2014-12-02 08:27:39 i wish i wouldnt need to do that but... 2014-12-02 08:27:41 :) 2014-12-02 08:27:42 ncopa, the init change i plan is http://sprunge.us/hMXJ 2014-12-02 08:28:02 perhaps you can have a staging repo? 2014-12-02 08:28:03 making fbsplash.cfg optional, and adding the forgotten 'mkfifo' that needs to be done to get working pipe 2014-12-02 08:28:14 (the abuild works for me) 2014-12-02 08:28:59 ScrumpyJack: yes becaue you have the file with the changes locally that was not included in the patch you sent 2014-12-02 08:29:02 git diff 2014-12-02 08:29:06 and you'll see 2014-12-02 08:29:22 so the Makefile.patch i got is different thatn the one you have 2014-12-02 08:30:04 fabled: im ok with that 2014-12-02 08:30:13 just a matter of taste but... 2014-12-02 08:30:35 I tend to prefer: if [ "$KOPT_splash" != "no" ] && [ -e $ALPINE_MNT/fbsplash.ppm ]; then 2014-12-02 08:30:44 which does exactly same thing 2014-12-02 08:30:59 but slightly more readable imho 2014-12-02 08:31:23 i must have got mixed up. I had forgotten something in the APKBUILD file so i reverted to before sending the previous patch 2014-12-02 08:32:44 that kind of things is what the checksum is supposed to catch ;) 2014-12-02 08:35:31 so why does abuild -r work for me on my filesystem in my aports repo? 2014-12-02 08:35:33 ncopa, pushed the init change 2014-12-02 08:35:57 ScrumpyJack: does git diff show any difference? 2014-12-02 08:36:20 fabled: i tested it here in qem 2014-12-02 08:36:21 qemu 2014-12-02 08:36:24 wroks 2014-12-02 08:36:31 yes, i tested ithere on rpi too 2014-12-02 08:36:32 works 2014-12-02 08:36:36 btw, i removed gtk and sdl support support from qemu 2014-12-02 08:36:55 i'm pretty happy on how 'professional' the Raspberry Pi looks now 2014-12-02 08:37:05 nice! 2014-12-02 08:37:11 1) blank screen for a second or so, when rpi is loading kernel/initramfs 2014-12-02 08:37:32 2) bootsplash (blinking cursors or messages are never seen if having blacklist=fbcon) 2014-12-02 08:37:42 3) omxplayer starts on top, and you get video 2014-12-02 08:37:52 veery nice looking 2014-12-02 08:37:59 the only bad thing is that boot takes about 60 seconds 2014-12-02 08:38:04 :) 2014-12-02 08:38:06 10-15 seconds is font-config 2014-12-02 08:38:11 ugh 2014-12-02 08:38:11 that is due fixed in their next release 2014-12-02 08:38:50 i suppose the way to boot it faster is to have it all in the initramfs... 2014-12-02 08:38:55 the rest of time hogs are likely init.d scripts, like dhcpcd and chrony, and other stuff being serialized 2014-12-02 08:39:25 nah. if it's all initramfs, or root squashfs - we still are tied with the MMC speed 2014-12-02 08:39:34 it does need to load everything 2014-12-02 08:39:39 faster MMC would help after that 2014-12-02 08:39:55 sounds like i should ask for a rpi :) 2014-12-02 08:40:03 but it think i'm happy with 45secs after fontconfig is fixed 2014-12-02 08:40:47 there's new model of it too 2014-12-02 08:41:35 bundle with all of it costs EUR 60 + tax here 2014-12-02 08:42:22 the plus model has hdmi, ethernet, 4x usb, 512mb ram and microsd 2014-12-02 08:42:39 I'm reading: http://mymediaexperience.com/raspberry-pi-xbmc-with-raspbmc/ 2014-12-02 08:43:19 ok 2014-12-02 08:46:10 git diff output is strange. Just says "standard input". i try to set a diff tool, but git difftool "is not a git command" 2014-12-02 08:57:09 git status show Your branch is ahead of 'origin/master' by 42 commits. 2014-12-02 09:05:19 diff.tool=/usr/bin/vimdiff 2014-12-02 09:05:47 git can be so frustrating when you don't know it very well :) 2014-12-02 09:13:20 ncopa: git diff shows nothing :( 2014-12-02 09:20:03 git diff --cached? 2014-12-02 09:25:01 empty 2014-12-02 09:35:46 ScrumpyJack: i dont know what went wrong then 2014-12-02 09:36:48 User-Agent: Alpine 2.11 (LNX 23 2013-08-11) 2014-12-02 09:37:04 it could be that your mail client change things 2014-12-02 09:37:12 i know it has wrapped lines in patches before 2014-12-02 09:37:24 so i does modify things 2014-12-02 09:40:49 yeah, alpine used to wrap lines. i've set it to flowtext since 2014-12-02 09:42:14 can you try this? http://st.ilet.to/frotz.patch 2014-12-02 09:42:27 ok, i found the issue :( :( 2014-12-02 09:43:32 ncopa, if that links works, it's my mail client 2014-12-02 09:51:29 ScrumpyJack: yes, it applies 2014-12-02 09:51:42 \o/ 2014-12-02 09:51:45 thanks 2014-12-02 09:52:18 stupid mail client 2014-12-02 09:52:31 gmail does the same 2014-12-02 09:53:30 since you are patch the makefile, why not fix the make install part too and use that? 2014-12-02 09:54:41 +-MAN_PREFIX = $(PREFIX) becomes + +-MAN_PREFIX = $(PREFIX) in my mail client :( 2014-12-02 09:56:46 have you tried the 'git send-email' client? 2014-12-02 09:57:06 fabled: are we ready for rc5? 2014-12-02 09:57:06 ncopa: I'm not clever enough to fix the make install part 2014-12-02 09:57:20 ncopa, we still have stuff that has not built 2014-12-02 09:57:22 forarm 2014-12-02 09:57:28 ok 2014-12-02 09:57:32 we wait some hours then 2014-12-02 09:57:47 ncopa: i don't have a local mta 2014-12-02 09:58:03 seems gtk-doc, usbutils and gnats are still broke (and the whole big dependency tree of them) 2014-12-02 09:58:11 you dont need local mta for git send-email 2014-12-02 09:58:11 i just fixed cpufrequtils 2014-12-02 09:58:18 not sure what's wrong with gtk-doc or usbutils 2014-12-02 09:58:23 but gnats is trying to sudo 2014-12-02 09:58:29 i fixed gnats 2014-12-02 09:58:46 ok 2014-12-02 09:59:11 samba seems to be compiling still for a while 2014-12-02 09:59:18 it's about 1/3th done 2014-12-02 09:59:23 i'll go eat meanwhile 2014-12-02 09:59:28 --smtp-server= ok 2014-12-02 09:59:42 http://git-scm.com/docs/git-send-email 2014-12-02 09:59:50 i'm there 2014-12-02 09:59:56 gtk-doc issue is funky. it works on edge builder 2014-12-02 09:59:57 there is an example how to configure it with gmail at the end 2014-12-02 10:00:15 in aports/.git/config 2014-12-02 10:00:20 usbutils seemed makefile / wrong include paths, so that might simpler to fix 2014-12-02 10:00:27 i'll try to get that fixed next 2014-12-02 10:00:31 and figure out gtk-doc 2014-12-02 10:00:31 add [sendemail] 2014-12-02 10:00:41 but now food 2014-12-02 10:00:43 ok 2014-12-02 10:01:10 speaking of git send email 2014-12-02 10:01:22 we should probaly have a separate mailing list for patches 2014-12-02 10:02:04 aports-patches@lists.alpinelinux.org or similar 2014-12-02 10:40:26 <^7heo> moin @ 2014-12-02 10:46:27 hi 2014-12-02 13:27:18 fabled: any idea what this error message is about? 2014-12-02 13:27:20 (8/51) Installing seafile-common (4.0.1-r0) 2014-12-02 13:27:20 ERROR: seafile-common-4.0.1-r0: No error information 2014-12-02 13:27:26 huh 2014-12-02 13:27:28 no 2014-12-02 13:27:48 i'm tracking / trying to workaround an annoying ffmpeg issue 2014-12-02 13:27:51 with deadline 2014-12-02 13:27:58 ok 2014-12-02 13:28:42 the bug is https://trac.ffmpeg.org/ticket/3175 2014-12-02 13:28:52 trying to figure out how to workaround it on alpine 2014-12-02 13:42:09 ok i think i got a workaround 2014-12-02 13:42:14 going to test it, brb. 2014-12-02 13:43:27 has the 'last' bug ever been figured out? 2014-12-02 13:43:41 which last bug? 2014-12-02 13:43:56 where when you use 'last' it says /dev/null/wtmp: no such file or directory 2014-12-02 13:44:09 its gnu coreutils bug 2014-12-02 13:44:20 oh 2014-12-02 13:44:24 and musl for defining wtmp stuff like that 2014-12-02 13:44:28 brb 2014-12-02 13:56:39 humpf both seafile and syncthing are broken 2014-12-02 13:57:58 syncthing is *huge* 2014-12-02 13:58:01 9.4M /usr/bin/syncthing 2014-12-02 13:58:01 4.3M /usr/bin/stevents 2014-12-02 13:58:01 3.9M /usr/bin/stindex 2014-12-02 13:59:31 ncopa-desktop:~$ syncthing 2014-12-02 13:59:31 14:58:53 FATAL: Invalid version string "v3.1.0_rc4+7-gc4ef7b3-dirty"; 2014-12-02 13:59:31 does not match regexp ^v\d+\.\d+\.\d+(-[a-z0-9]+)*(\+\d+-g[0-9a-f]+)?(-dirty)?$ 2014-12-02 13:59:43 its picking up the git version from aports 2014-12-02 14:05:47 the regex does match the git version 2014-12-02 14:06:14 but _rc4 does not match (-[a-z0-9]+)* 2014-12-02 14:18:24 nsz: syncthing picks up git describe from our buildscrips repo 2014-12-02 14:18:37 so the version number is picks up build time is wrong 2014-12-02 14:40:22 ncopa: that syncthing thingy sounds like a really cute bug 2014-12-02 14:51:22 barthalion: i fixed it already 2014-12-02 14:52:52 ScrumpyJack, ping 2014-12-02 15:07:35 alacerda: pong 2014-12-02 15:09:44 did you see the pvt i sent you? 2014-12-02 15:11:09 about LXDE? 2014-12-02 15:11:38 sorry, i just saw your private message 2014-12-02 15:13:06 do you want me to change the feature request to LXQt? 2014-12-02 15:14:16 ScrumpyJack, would be better 2014-12-02 15:14:17 ;) 2014-12-02 15:14:27 if you don't mind 2014-12-02 15:14:29 will do 2014-12-02 15:49:12 our first hundred followers on twitter \0/ 2014-12-02 18:16:58 rnalrd / fabled / ncopa : can we remove the note from the beginning of http://wiki.alpinelinux.org/wiki/How-To_Alpine_Wall ? 2014-12-02 18:17:22 "This material is work-in-progress ..." 2014-12-02 18:57:44 <^7heo> tdtrask: this is a wiki 2014-12-02 18:57:47 <^7heo> tdtrask: just edit it. 2014-12-02 18:58:05 <^7heo> and post on the ML that you did, so if someone has a problem with what you did 2014-12-02 18:58:18 <^7heo> he/she will restore the notice. 2014-12-02 19:02:08 meh 2014-12-02 19:03:12 tdtrask: I don't remember by login and password, sorry 2014-12-02 19:03:25 <^7heo> s/b/m/ 2014-12-02 19:06:53 ACTION is capable of removing the note, but wanted t ocheck with those who would know before doing so 2014-12-02 19:07:02 tdtrask: done. 2014-12-02 19:07:16 ACTION doesn't have the faith of ^7heo that those responsible would notice 2014-12-02 19:07:20 :) 2014-12-02 19:07:44 there is a log of wiki changes that folks watch, and can revert 2014-12-02 19:08:37 (obvioulsy) 2014-12-02 19:08:54 ACTION can't spell 2014-12-02 19:11:48 why gnome-session doesn't exist in repo anymore ( since v3.0 )? 2014-12-02 19:16:42 tdtrask, do you know sth about it ^^ ? 2014-12-02 19:18:06 up to v2.7 gnome-session was in the repo. On v3.0 it is not 2014-12-02 19:18:30 <^7heo> alacerda: 'cause gnome sucks 2014-12-02 19:19:52 I'd express my concerns about who else sucks here, but I disgress 2014-12-02 19:20:01 ^7heo, thanks for your opnion ;) 2014-12-02 19:20:14 alacerda: we dropped it as there was no way we could build newer release 2014-12-02 19:20:31 I am trying to fix this: http://wiki.alpinelinux.org/wiki/Gnome_Setup 2014-12-02 19:20:35 <^7heo> alacerda: it's a fact :P 2014-12-02 19:20:50 alacerda: the best way to fix it is to package MATE 2014-12-02 19:20:59 if we follow those steps, we'll have a xfce environment and not a gnome 2014-12-02 19:21:27 <^7heo> I will never understand the need for things such as "desktop environments" 2014-12-02 19:21:42 tmux is all we need 2014-12-02 19:21:45 and vi 2014-12-02 19:21:45 <^7heo> it's like if the unix philosophy was actively underemined by users... 2014-12-02 19:21:56 <^7heo> barthalion: I'm not saying that. 2014-12-02 19:21:59 I do 2014-12-02 19:22:01 <^7heo> you might want to browse the web. 2014-12-02 19:22:06 <^7heo> or play games 2014-12-02 19:22:07 with w3m 2014-12-02 19:22:08 <^7heo> stuff like that 2014-12-02 19:22:14 <^7heo> or watch videos 2014-12-02 19:22:17 <^7heo> etc. 2014-12-02 19:22:20 <^7heo> but anyway 2014-12-02 19:22:54 <^7heo> that's not a reason for using a fully bloated Desktop Env. 2014-12-02 19:23:33 ^7heo, gracias! good to know your opinion. :) 2014-12-02 19:23:56 alacerda: but going back to the topic… ask someone or pay him for packaging mate 2014-12-02 19:24:32 <^7heo> alacerda: stop trolling =/ 2014-12-02 19:24:38 <^7heo> alacerda: it was fun the first time. 2014-12-02 19:24:46 we could probably build gnome 3.14 with gentoo/freebsd patches, but we really have no manpower for that 2014-12-02 19:24:59 <^7heo> and yes it is my opinion. 2014-12-02 19:25:20 <^7heo> but you can drop the sarcasm and the passive aggressive smiley, that's not helping. 2014-12-02 19:25:44 alacerda: and I'm hardly finding time for Arch, not to mention Alpine, unfortunately 2014-12-02 19:27:08 barthalion, tks for your helping answer about mate 2014-12-02 19:54:11 hmm, that how-to awall doesn't read very well, tdtrask, why did you want to remove the draft? 2014-12-02 19:55:50 <^7heo> barthalion: focus on arch, we don't need arch people sabotaging alpine. 2014-12-02 19:57:14 ^7heo: you are boring 2014-12-02 19:57:17 I hope you know that 2014-12-02 19:57:19 :) 2014-12-02 19:58:11 <^7heo> and your english sucks. 2014-12-02 19:58:23 <^7heo> or you need to grow a pair. 2014-12-02 19:58:30 <^7heo> it's either/or. 2014-12-02 19:58:40 <^7heo> now, I wasn't hating Archlinux, initially. 2014-12-02 19:59:01 <^7heo> maybe if the project was featuring real devs I still wouldn't. 2014-12-02 19:59:05 ^7heo: I really cannot care less about your opinion 2014-12-02 19:59:15 so I'll just make use of /ignore 2014-12-02 19:59:34 <^7heo> suit yourself. 2014-12-02 20:00:18 <^7heo> at least I can bitch about arch without having trolls reacting. 2014-12-02 20:00:29 <^7heo> sorry I meant arch devs. 2014-12-02 20:00:49 <^7heo> I always mix those two terms, I dunno why. 2014-12-02 20:24:51 ScrumpyJack: that's the best doc we have 2014-12-02 20:25:08 and there was a request on the mailing list for documentation 2014-12-03 07:19:07 morning 2014-12-03 09:23:38 <^7heo> moin 2014-12-03 09:25:39 <^7heo> barthalion: hey 2014-12-03 09:26:17 <^7heo> barthalion: I wanted to apologize for my behavior towards you, yesterday. 2014-12-03 09:26:41 <^7heo> barthalion: I have been too harsh, and you mentionning arch isn't a valid reason for such a behavior. 2014-12-03 09:26:48 <^7heo> barthalion: so, sorry. I won't do it again. 2014-12-03 09:27:24 <^7heo> and if you wish to respect me, please REFRAIN from mentionning arch whenever you possibly can. 2014-12-03 09:41:27 hmmh 2014-12-03 09:42:05 <^7heo> yeah, I know, he probably ignored me already; so that's useless. 2014-12-03 09:42:31 ^7heo: i think he is away 2014-12-03 10:51:50 ncopa, for the new website: http://sprunge.us/UYaL 2014-12-03 10:54:15 <^7heo> ah yeah, that's another way of submitting patches... 2014-12-03 10:55:09 works great as long someone is available to deal with the patch instantly 2014-12-03 10:55:36 applied 2014-12-03 10:55:38 thanks! :) 2014-12-03 10:55:42 ncopa, thanks 2014-12-03 10:56:10 this is one thing i like with new site... 2014-12-03 10:56:16 people can send me patches :) 2014-12-03 10:56:19 oh reminds me 2014-12-03 10:56:41 i should move the git repo 2014-12-03 10:57:42 ncopa, i think this "send-patch" style would work well for documentations also 2014-12-03 11:00:54 better than wiki? 2014-12-03 11:01:02 actually been thinking of that 2014-12-03 11:01:28 having official documentation as asciidoc 2014-12-03 11:01:33 I like asciiDoc + Git 2014-12-03 11:01:37 yeah! :) 2014-12-03 11:02:16 +1 2014-12-03 11:02:19 I like this idea 2014-12-03 11:07:19 +1 for asciidoc 2014-12-03 11:24:35 <^7heo> alacerda: wait, aren't the dude/gal I taunted yesterday, too? 2014-12-03 11:25:10 <^7heo> ah no, you were sarcastic. 2014-12-03 11:25:13 <^7heo> my bad. 2014-12-03 11:25:34 <^7heo> on the side of solutions that already work: markdown + git. 2014-12-03 11:25:39 <^7heo> it's called ikiwiki. 2014-12-03 11:25:59 <^7heo> it's written in perl (yes, I know), but features some cgi thing if you need to allow web edition. 2014-12-03 11:26:20 <^7heo> it's MUCH better than any other wiki engine for the fact that it uses git. 2014-12-03 11:27:14 <^7heo> if you don't need the web edition, you can always go for a C markdown/asciidoc parser 2014-12-03 11:27:34 <^7heo> (invoked via a git hook, on a public repo - done) 2014-12-03 11:29:33 "Ikiwiki is a wiki compiler. It converts wiki pages into HTML pages suitable for publishing on a website." seems to be a nice tool! :) 2014-12-03 11:29:53 I didn't know it 2014-12-03 11:30:09 <^7heo> it is okay 2014-12-03 11:30:14 <^7heo> minus the perl aspect. 2014-12-03 11:30:28 <^7heo> (which some would argue to be a plus, but not me) 2014-12-03 11:31:52 well... it is another option to put in the list/to discuss! 2014-12-03 11:33:37 <^7heo> imho it should be used if web-edits are required; it's already existing and can be deployed. 2014-12-03 17:16:59 ^7heo: apology accepted; deal with the fact that I develop Alpine on Arch, and anything that I push here is tested first in your beloved distribution with, for better or worse, bigger userbase 2014-12-03 18:05:28 <^7heo> barthalion: whatever mistake you do isn't my problem for now, and if it is, either you will live or I will. 2014-12-03 18:06:01 <^7heo> but mistakes aren't the real problem imho 2014-12-03 18:06:27 <^7heo> the problem is arrogance and ignorance combined. 2014-12-03 18:06:34 <^7heo> so as long as you're not both I'm fine. 2014-12-03 18:07:02 <^7heo> (and by that I mean "being arrogant and ignorant", not "being arrogance and ignorance", ofc) 2014-12-03 18:07:19 <^7heo> omg, I meant leave not live. 2014-12-03 18:07:35 <^7heo> sorry, I tend to write engrish when I'm provoked. 2014-12-04 06:55:38 annoying. arm builder is so slow. 2014-12-04 06:55:42 still waiting for it to catch up 2014-12-04 07:00:56 morning 2014-12-04 07:00:57 yeah 2014-12-04 07:01:10 i wrote a requisition for 2 wandboards 2014-12-04 07:02:50 nice 2014-12-04 07:03:01 i think the case now is that both edge, and 3.1-stable are catching up 2014-12-04 07:03:08 and it can cause swapping at times 2014-12-04 07:03:18 the poor box has only 2 gigs of ram 2014-12-04 07:03:29 and it's building lots of c++ stuff 2014-12-04 07:04:38 it's *still* building webkit on 3.1-stable 2014-12-04 07:04:46 yes 2014-12-04 07:05:00 i remembered it's especially slow 2014-12-04 07:05:07 but 24+h is ridiculous 2014-12-04 07:05:59 it does not help that we build it twice 2014-12-04 07:06:16 one for gtk3 and one for gtk2 2014-12-04 07:38:31 ncopa, i'll upgrade musl too; cherry-pick the important bug-fixes from git with few new things that improve compatibility 2014-12-04 07:54:47 <^7heo> http://arm.slackware.com/buildfarm/ 2014-12-04 07:54:56 <^7heo> solution 1: ask them 2014-12-04 07:55:04 <^7heo> solution 2: do one :) 2014-12-04 11:37:27 I added dconf to makedepends but abuild -r says: 2014-12-04 11:37:28 checking for DCONF... no 2014-12-04 11:37:28 checking for DCONF... no 2014-12-04 11:37:28 configure: error: Package requirements (dconf >= 0.10.0) were not met: 2014-12-04 11:37:46 Package dconf was not found in the pkg-config search path. 2014-12-04 11:37:50 o.0 2014-12-04 11:37:54 any idea? 2014-12-04 12:30:41 alacerda: dconf-dev 2014-12-04 12:43:04 barthalion, thanks again ;) 2014-12-04 13:36:25 hmm, I have two patches in a abuild, created the same way, for two different files. one works, the other doesn't 2014-12-04 14:00:43 arm builder is still working on webkitgtk 2014-12-04 14:01:11 yeah, edge is almost done catching up; there's still few things left though 2014-12-04 14:04:31 hmm, I've been making a prayer apk and i see that someone has pushed one to testing, even though i assigned it, and i didn't see a post on mailing list. 2014-12-04 14:16:38 while i appreciate that someone else (faster than me) has pushed a patch for this package, it would be very helpful if patches of new aports were announced on the mailing list and the feature request updated as required. (I, for one, have very little free time that I can spend on alpine - which could be better spent by *not* duplicating work) 2014-12-04 14:30:44 <^7heo> ncopa: since yesterday? 2014-12-04 14:43:04 yes 2014-12-04 14:43:31 <^7heo> =/ 2014-12-04 14:57:41 can anyone with inkscape/svg skills give me a hand? 2014-12-04 14:57:50 I have this: http://alpinelinux.org/alpinelinux-logo-icon.svg 2014-12-04 14:57:57 and would like it as .png 2014-12-04 14:58:11 for wiki, git and forums sites 2014-12-04 14:58:31 i have this too: http://alpinelinux.org/alpinelinux-logo.svg 2014-12-04 14:58:54 so generate png images in proper sizes 2014-12-04 14:59:01 from the above svgs 2014-12-04 14:59:17 <^7heo> and you don't have inkscape or you don't know how to do it with inkscape? 2014-12-04 14:59:25 i know how to do it 2014-12-04 14:59:29 i just want save time... 2014-12-04 14:59:30 <^7heo> imho even imagemagik can do it 2014-12-04 14:59:32 <^7heo> like with convert 2014-12-04 14:59:36 <^7heo> lemme try 2014-12-04 15:00:00 also find out which might be the proper size for each site 2014-12-04 15:00:10 <^7heo> yeah 2014-12-04 15:00:12 <^7heo> works 2014-12-04 15:00:22 <^7heo> convert alpinelinux-logo.svg alpinelinux-logo.png 2014-12-04 15:00:23 <^7heo> done. 2014-12-04 15:00:42 <^7heo> $ apk info -W $(which convert) 2014-12-04 15:00:42 <^7heo> /usr/bin/convert is owned by imagemagick-6.9.0.0-r0 2014-12-04 15:06:37 <^7heo> or if you have inkscape you can use: 2014-12-04 15:07:09 <^7heo> inkscape -z -e alpinelinux-logo.png -w -h alpinelinux-logo.svg 2014-12-04 15:07:27 <^7heo> ('cause it seems that convert FIRST converts to png and THEN scales) 2014-12-04 15:07:51 <^7heo> (with convert you need to increase the DPIs first I believe) 2014-12-04 15:33:21 i'd be happy if you could just send me th converted .png files in an email saying which png goes where... 2014-12-04 15:33:34 oh 2014-12-04 15:33:44 maybe for bugs.alpinelinux.org too if needed 2014-12-04 16:04:43 /msg ncopa i have a couple a free hours tonight. anything i can do? 2014-12-04 16:04:53 /msg ncopa: 2014-12-04 16:04:56 arg 2014-12-04 16:11:19 <^7heo> ScrumpyJack: fail 2014-12-04 16:11:20 <^7heo> :) 2014-12-04 16:26:49 <^7heo> docker in alpine is far from perfect 2014-12-04 16:26:56 fix it 2014-12-04 16:27:09 <^7heo> no time. 2014-12-04 16:27:10 ScrumpyJack: sorry, I'll pay more attention to the bug tracker next time 2014-12-04 16:27:24 <^7heo> if I had time, I wouldn't report, I would fix. 2014-12-04 16:27:45 "far from perfect" is far from accurate report 2014-12-04 16:27:53 <^7heo> barthalion: so that's how you react? When people sincerely apologize, you make them regret? 2014-12-04 16:28:14 ^7heo: bring your psychoanalysis somewhere else, I have no time for it 2014-12-04 16:28:19 <^7heo> and you become aggressive since they have stepped from aggressive to trying to respect you? 2014-12-04 16:28:32 <^7heo> yeah oay. 2014-12-04 16:28:34 <^7heo> okay. 2014-12-04 16:28:38 <^7heo> typical archlinux dev. 2014-12-04 16:28:49 ScrumpyJack: the guy who requested it messaged me directly and I had a spare 15 minutes, didn't know about you working on it 2014-12-04 16:29:17 <^7heo> people like you make me think twice before apologizing, but I'll still apologize: I don't want to be as awful as you are. 2014-12-04 16:29:18 ScrumpyJack: could you verify if it actually works? 2014-12-04 16:29:31 <^7heo> I'm better than that. 2014-12-04 16:29:49 <^7heo> anyway, docker is far from perfect, and I reported it partially already. 2014-12-04 16:30:52 <^7heo> so far, two issues that I can think of: flags are matching the man description, but aren't the same as the ones present in other distros (different versions?); and removing images doesn't remove them all. 2014-12-04 16:31:09 ScrumpyJack: the code was awfully old and unmaintained enough to be dropped from debian, so unless it works flawlessly I wouldn't move it to main 2014-12-04 16:37:33 <^7heo> okay people, I guess that archlinux has plaggued alpine way faster than I could have imagined. 2014-12-04 16:37:37 <^7heo> laters 2014-12-04 17:49:57 barthalion: sure, I'll give it a whirl. thanks for taking the time to port it. it took me *much* longer than 15 minutes to patch it to get it to compile. 2014-12-04 17:51:03 i looked for arch patches but they had nothing so i started from scratch. did you find anything? 2014-12-04 17:56:33 I searched for a solution for this problem: http://hastebin.com/jaxowozeko.vhdl 2014-12-04 17:56:57 and found ncopa's answer: http://hastebin.com/balipeyoda.xml 2014-12-04 17:57:17 but I tried with no results 2014-12-04 17:57:19 =/ 2014-12-04 18:01:19 ScrumpyJack: I found debian patches 2014-12-04 18:02:03 alacerda: as the last command in package(): find "$pkgdir" -name \*.la -delete 2014-12-04 18:02:19 ScrumpyJack: I don't think we ever had prayer in Arch 2014-12-04 18:02:35 ScrumpyJack: ah, debian didn't drop it, just patched a lot 2014-12-04 18:31:58 barthalion, didnt work 2014-12-04 18:32:54 i've already tried to make CFLAGS=libtool install 2014-12-04 18:33:06 but didnt work either 2014-12-04 18:59:46 alacerda: show me your APKBUILD then 2014-12-04 19:00:19 barthalion, http://sprunge.us/EBaX 2014-12-04 19:00:52 just remembering: i made lots of changes... the line you sent isn't there anymore 2014-12-04 19:03:07 and barthalion i noticed that I made a mistake 2014-12-04 19:03:18 try again with it if you can 2014-12-04 19:03:32 that command you asked me tu put on package() 2014-12-04 19:03:34 ... 2014-12-04 19:03:38 i put on build() 2014-12-04 19:03:38 yeah 2014-12-04 19:03:40 duhhh 2014-12-04 19:03:47 let me try again 2014-12-04 19:04:03 before the make isn't? 2014-12-04 19:04:07 no, after 2014-12-04 19:04:15 o 2014-12-04 19:04:16 ok 2014-12-04 19:04:35 you want to get rid of la files after they hit $pkgdir 2014-12-04 19:04:42 before make install they are not there yet 2014-12-04 19:05:06 I thought they were put there after "make" 2014-12-04 19:05:25 yes, that's why you run find after make install 2014-12-04 19:05:55 make != make install; no??? 2014-12-04 19:06:01 yes sir 2014-12-04 19:06:06 hehe 2014-12-04 19:06:09 make builds, make install installs 2014-12-04 19:06:23 see poppler's package() for an example 2014-12-04 19:07:27 wow... it is there also! 2014-12-04 19:07:28 :) 2014-12-04 19:07:52 this is something we should be doing in abuild by default, maybe tommorow I'll send a patch 2014-12-04 19:09:41 barthalion, thanks 2014-12-04 19:09:42 ;) 2014-12-04 19:09:45 worked 2014-12-04 19:10:00 you're welcome 2014-12-04 19:10:04 and it'll be great if abuild start doing this 2014-12-05 08:10:52 ncopa: http://sprunge.us/fMKV does it make any sense to you? 2014-12-05 08:13:33 it does make sense 2014-12-05 08:13:51 i tried to avoid doing things behind the back of user 2014-12-05 08:13:52 but 2014-12-05 08:14:15 i think that we reached the point where more conveninent maintenance is wanted 2014-12-05 08:14:27 we have to wait til after v3.1 though 2014-12-05 08:14:41 and 2014-12-05 08:14:53 i think i'd like to rething the entire APKBUILD concept... 2014-12-05 08:15:03 rethink* 2014-12-05 08:25:12 I'm unsure if libs() is the best place for that 2014-12-05 08:25:20 something like post_package would be handy 2014-12-05 08:25:45 do we control the order of creating packages and subpackages? 2014-12-05 08:25:52 yes 2014-12-05 08:26:10 subpackages="first second third" 2014-12-05 08:26:58 I need to think about it more, as my patch doesn't apply to all use cases 2014-12-05 08:27:11 but that's later, going back to work 2014-12-05 08:27:47 i would actually like to replace the APKBUILD format with soemthign else 2014-12-05 08:28:01 there are more issues that is not easy to resolve 2014-12-05 08:37:13 what I want to achieve is deleting *.la files in main package and each subpackage 2014-12-05 08:39:50 understood 2014-12-05 08:39:59 lets look at that later 2014-12-05 08:40:02 after v3.1. is out 2014-12-05 08:42:08 looks like 3.1-stable/arm builder has quite a long way to go 2014-12-05 08:45:44 :-/ 2014-12-05 10:20:02 ncopa ping.. 2014-12-05 10:30:54 jarruda: pong 2014-12-05 10:33:13 ncopa... i got AL 3.1.0_rc4 with FreeSWITCH version: 1.4.13 running in my lab.. 2014-12-05 10:33:21 cool 2014-12-05 10:34:35 and I can consistently replicate a scenario where freeswitch stucks after a call... 2014-12-05 10:35:29 ok 2014-12-05 10:35:50 the problem is there is no clues from the logs or FS debug messages... 2014-12-05 10:35:56 ok 2014-12-05 10:36:12 have you tried get help from FS devs? 2014-12-05 10:36:26 or reported how to reproduce it on FS jira? 2014-12-05 10:36:50 the process freezes and will not even restart... (kill -9 is the only way I manage to stop it) 2014-12-05 10:38:04 hm 2014-12-05 10:38:10 there is freeswitch 1.4.14 out 2014-12-05 10:38:41 check if you find soemthing that looks relevant in changelog 2014-12-05 10:40:02 we could also try build it with debugging symbols enabled and then send it the signal to generate a core dump 2014-12-05 10:40:09 or maybe even run it directly in gdb 2014-12-05 10:40:29 so we can analyze what is going on 2014-12-05 10:40:40 to analyze where it gets stuck 2014-12-05 10:41:50 \o/ 2014-12-05 10:41:52 whee! 2014-12-05 11:16:58 ncopa nothing from fs 1.4.14 changelog jumps to be related with this (just to add information, the same behaviour is been observed in AL 3.0.6 where the FS version is 1.4.6) 2014-12-05 11:28:05 clandmeter and ncopa, what do you think of creating another mail list, besides "alpine-devel", for website devel? Maybe sth like "alpine-site-devel" ( it would be used for the documentation also - when/if we migrate to asciiDoc ) 2014-12-05 13:54:21 hum... 2014-12-05 13:54:38 http://bld1.alpinelinux.org/buildlogs/build-edge-x86/main/gradm/gradm-3.0.201408301734-r0.log 2014-12-05 14:17:26 ScrumpyJack: can you install alpine disk install with 70MB? 2014-12-05 14:17:36 i thought kernel was bigger than that 2014-12-05 17:32:06 ping barthalion 2014-12-05 17:32:50 are you familiar with this error? http://hastebin.com/uxijakukay.coffee 2014-12-05 17:39:20 alacerda: pass -D_GNU_SOURCE to gcc 2014-12-05 17:59:21 thanks barthalion 2014-12-05 18:05:06 np 2014-12-05 18:33:51 alacerda: http://lists.alpinelinux.org/alpine-devel/4527.html 2014-12-05 18:34:16 barthalion, \o/ tkx... 2014-12-05 18:34:39 i had to do this for a lot of packages i am porting here 2014-12-05 18:37:28 sent v2 with fixed summary 2014-12-05 18:37:44 I forgot to follow ncopa's commit pattern 2014-12-07 07:23:00 Hi , i want to use a udevfree fluxbox-gtkbased (non-alpine) desktop..mdev+openrc .boots fine .wlan.usb ok,console keyboard ok.startx ok (intel),but no keyboard no mouse. tried alot of xorg.confs .could you explain the right xorg.conf for a nonudevsysytem? thanks 2014-12-07 07:26:11 ...what i dont get is :if there is a "driver" working during console,how do i identidy that and adopt it for xorg.conf... 2014-12-07 07:26:37 sorry -> "identify" 2014-12-07 12:04:59 I'll be blocking buildbots for a while today 2014-12-07 12:05:15 I'm going to push new ffmpeg and fix build errors in the fly 2014-12-07 12:05:42 3.1 is branched out already? 2014-12-07 12:06:52 nope 2014-12-07 12:07:02 ok, I'll wait for it to introduce breakages 2014-12-07 12:07:26 not to*, of course :D 2014-12-08 08:54:23 will AL v3.1 have fixed pgks release or updated pkgs rolled over it, like v3.1.1 ? 2014-12-08 08:57:05 yes 2014-12-08 08:59:13 will it be difficult to matain v3.1 and v3..1.1 separately ? 2014-12-08 08:59:22 maintain* 2014-12-08 09:02:33 we branch 3.1-stable 2014-12-08 09:02:54 so v3.1 will be maintained lie v3.0 2014-12-08 09:03:43 and v3.1.1 ? 2014-12-08 09:04:29 will be in v3.1 branch 2014-12-08 09:04:50 we will have v3.1.x releases 2014-12-08 09:04:56 not v3.1.1.x 2014-12-08 09:05:03 why would we? 2014-12-08 09:05:42 I think v3.1.x should suffice 2014-12-08 09:06:42 can both the updated pkgs and old pkgs available via nl.a.o 2014-12-08 09:06:48 like http://nl.alpinelinux.org/alpine/v3.1 and http://nl.alpinelinux.org/alpine/v3.1.1 ? 2014-12-08 09:07:31 one can just add http://nl.alpinelinux.org/alpine/v3.1.1 to /etc/apk/world 2014-12-08 09:07:51 rather append 2014-12-08 09:17:31 also possible is http://nl.alpinelinux.org/alpine/v3.1/updates/1 2014-12-08 09:17:38 you should replace v3.0 with v3.1 in /etc/apk/repositories 2014-12-08 09:17:53 http://nl.alpinelinux.org/alpine/v3.1 is the updates 2014-12-08 09:19:15 not refering to 3.0.x 2014-12-08 09:19:50 just next release i.e 3.1 and its updates 2014-12-08 09:23:28 though by same logic it is possible to replace http://nl.alpinelinux.org/alpine/v3.1 to http://nl.alpinelinux.org/alpine/v3.1.1 in /etc/apk/repositories 2014-12-08 09:24:12 but then file system structure in nl.a.o would be different 2014-12-08 09:25:07 re: mistake my posting above /etc/apk/world should be read as /etc/apk/repositories 2014-12-08 09:29:47 no, i will not set up a builder for v3.1.1.x releases 2014-12-08 09:30:10 we only have v3.x branches 2014-12-08 09:30:48 http://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases 2014-12-08 09:34:31 why the need to set separate builder ? 2014-12-08 09:35:21 and I am not thinking of v3.1.1.x 2014-12-08 09:36:31 or is http://nl.alpinelinux.org/alpine/v3.1/updates/1 better ? 2014-12-08 09:47:14 wat 2014-12-08 09:49:37 possible not to just set /updates/ ? 2014-12-08 09:50:01 for pkgs output directory 2014-12-08 09:50:13 but what's the point? 2014-12-08 09:50:37 updates just hit /v3.1/ 2014-12-08 09:50:54 1. roll back , 2, regression in newer pgks 2014-12-08 09:50:58 v3.1.x versioning is more about installers 2014-12-08 09:51:31 when there is a regression you should report it here/alpine-devel/bugtracker and the package will be fixed or reverted 2014-12-08 09:51:50 the only things we push to stable releases are bug and security fixes 2014-12-08 09:51:53 yes, but should be site be down till then ? 2014-12-08 09:52:06 of just compile yourself 2014-12-08 09:52:23 you should keep your own cache if you take life seriously :p 2014-12-08 09:52:36 so no overall sense of stability 2014-12-08 09:52:48 ah... cache it, got it 2014-12-08 09:54:49 that means stable releases are also "updated rolling releases" 2014-12-08 09:56:55 which also means you cannot tag any release as "stable" 2014-12-08 10:00:35 it's not true either 2014-12-08 10:00:39 they are not rolling 2014-12-08 10:00:46 btw, how do you cache it, any strategy 2014-12-08 10:00:58 you have specified set of versions that aren't changed 2014-12-08 10:01:03 setup-apkcache ;) 2014-12-08 10:01:39 and what happens after apk update, apk upgrade ? 2014-12-08 10:01:57 can rollback after that ? 2014-12-08 10:02:41 so how does one define "rolling" in os context ? 2014-12-08 10:33:42 vkrishn: you cannot really rollback 2014-12-08 10:33:56 well you can if you have set up the apk cache 2014-12-08 10:34:04 and then you'll have to do something like: 2014-12-08 10:34:26 apk add "broken-package<$version" 2014-12-08 10:34:34 where version is the broken version 2014-12-08 10:35:00 then you tell apk to only instaly something older than $version 2014-12-08 10:36:03 or during apk upgrade capture the list of files updated and during rollback do apk ? 2014-12-08 10:36:17 apk del 2014-12-08 10:36:31 then use old pkg repo to apk add 2014-12-08 10:37:12 but wait , there is no old repo !! 2014-12-08 10:38:16 unless we start in v3.1 2014-12-08 10:39:39 vkrishn: do you think we should vreate a new repo for every package we update? 2014-12-08 10:39:59 no tag like v3.1.1 , v.3.1.2 2014-12-08 10:40:51 so every time user wants install security updates he needs to edit repositories file? 2014-12-08 10:41:29 depends on how you create the update repo directory 2014-12-08 10:41:48 with just updated pkgs, append the new repo 2014-12-08 10:42:38 I can suggest 2 methods if opting for full fledged update repo directory 2014-12-08 10:45:07 also possible is logic to process http://nl.alpinelinux.org/alpine/v3.1/updates directory 2014-12-08 10:45:22 so no need to edit repository file 2014-12-08 10:46:36 it depends how you adapt in/to existing code or infra 2014-12-08 10:48:36 one that involves less maintainance and better output could be best 2014-12-08 10:58:21 and yes, this proces is for version releases only, not edge 2014-12-08 10:58:48 for edge kernel updates could be one option 2014-12-08 11:00:31 but yet to think more on edge 2014-12-08 11:34:06 I see ffmpeg 2.5, why still at 2.3.5 ? due to ABI change ? 2014-12-08 11:34:47 3.1 will be out soon 2014-12-08 11:34:58 no reason to introduce potential breakage 2014-12-08 11:35:07 I'll push if after 3.1 is branched 2014-12-08 11:35:23 yes but 2.3.5 was done 24 hrs back 2014-12-08 11:35:38 when 2.4.4 existed 2014-12-08 11:35:58 so what? 2.4 isn't ABI compatible with 2.3 2014-12-08 11:36:08 neither is 2.5 2014-12-08 11:36:18 ok, that's what wanting to know, thanks 2014-12-08 11:36:39 although 2.5 is compatible with 2.4, so it will make things simpler 2014-12-08 11:36:57 so 2.4.4 would have introduced a chain update 2014-12-08 11:45:21 not a chain update, but a couple of possibly incompatible packages to rebuild 2014-12-08 11:45:52 ok, kinda then heavy at this point 2014-12-08 12:04:23 main reason to not update ffmpeg is that current vlc does not support it 2014-12-08 12:04:34 2.4 does not only break ABI 2014-12-08 12:04:38 it also breaks API 2014-12-08 12:05:25 ncopa: that's not entirely true 2014-12-08 12:05:31 ncopa: vlc builds fine with new vlc 2014-12-08 12:05:44 vlc configure script complained 2014-12-08 12:05:50 yeah, but that's all 2014-12-08 12:05:53 ok 2014-12-08 12:06:01 misunderstanding then 2014-12-08 12:06:02 I fixed that in configure.ac, ran autoreconf and everything built fine 2014-12-08 12:06:08 heh 2014-12-08 12:06:16 ok 2014-12-08 12:06:19 I was surprised too 2014-12-08 12:06:34 total pkgs with ffmpeg-dev = 12 2014-12-08 12:07:13 in main/ 2014-12-08 12:07:22 so we upgrade it? 2014-12-08 12:07:30 is it worth another day delay? 2014-12-08 12:07:35 nope 2014-12-08 12:07:40 it is not 2014-12-08 12:09:38 just got curious as 2.4.4 existed while 2.3.5 got in edge 2014-12-08 12:11:40 why there is no framework pkg for lua ? 2014-12-08 12:12:28 eg http://www.keplerproject.org/ , Xavante 2014-12-08 12:15:37 https://github.com/hishamhm/lua-compat-5.2 , could this be useful as we have v5.1 ? 2014-12-08 12:17:28 possibly 2014-12-08 12:17:49 i have a mod that i havent been able to port to 5.2 2014-12-08 12:17:56 lua-subprocess iirc 2014-12-08 12:21:11 also any nice documented lua framework pkg would be nice 2014-12-08 12:23:47 \o/ 2014-12-08 12:25:06 what !!, I don't see any git tag ? 2014-12-08 12:26:20 i could not tag unless build-3-1-armhf built it all 2014-12-08 12:26:57 https://github.com/hishamhm/lua-compat-5.2 - is Lua-5.2-style APIs for Lua 5.1 2014-12-08 12:28:14 also nice, https://github.com/hishamhm/luaposix 2014-12-08 12:28:30 ahmm, seems nice 2014-12-08 12:29:08 https://github.com/luaposix/luaposix 2014-12-08 13:53:40 \o/ 2014-12-08 13:53:44 http://nl.alpinelinux.org/alpine/v3.1/releases/armhf/ 2014-12-08 13:53:55 rpi releases was generated 2014-12-08 13:55:23 nice !! 2014-12-08 13:56:27 time for a short relax in coffee bar ? 2014-12-08 13:57:03 ncopa: can we have an iso of armhf that is not rpi specific? 2014-12-08 13:57:37 ScrumpyJack: not for v3.1 i think 2014-12-08 13:57:50 i have ordered 2 wandboards 2014-12-08 13:57:59 ScrumpyJack, armhf is kinda special; you need bootloader / kernel for the specific board 2014-12-08 13:58:08 i run alpine on my wandboard builder 2014-12-08 13:58:17 but we have no release images for that yet 2014-12-08 13:58:26 will try to add them soon though 2014-12-08 13:58:31 so ncopa can get additional builders up 2014-12-08 13:58:38 :) 2014-12-08 13:59:01 fabled: there are a bunch of security bugs in binutils 2014-12-08 13:59:09 i wonder if i should just upgrade to binutils-2.5 2014-12-08 13:59:16 but i think its scary... 2014-12-08 14:01:07 http://seclists.org/oss-sec/2014/q4/449 2014-12-08 14:04:09 fabled: i was thinking of an iso for qemu 2014-12-08 14:04:43 ScrumpyJack, you'd need qemu to be configured for some specific arm board 2014-12-08 14:04:57 i see 2014-12-08 14:07:50 so you couldnt just use an iso to boot a riscpc 2014-12-08 15:47:18 thx :-) 2014-12-08 18:02:19 raspberry pi release, neato 2014-12-08 18:02:29 makes me want to get one now but idk what i would use it for :p 2014-12-09 06:56:36 ncopa, i tested _rc5 on rpi and it's good (all fbcon / blacklist / splash stuff included) 2014-12-09 06:56:56 though, the dmvpn gre/gso issue remains that was supposed to be fixed by the netdev patch we applied 2014-12-09 07:12:12 there is a new kernel out 2014-12-09 07:12:15 3.14.26 2014-12-09 07:12:50 fabled: so i simply remove the gre/gro patch? 2014-12-09 07:13:04 does it conflict? 2014-12-09 07:13:07 i'm still thinking 2014-12-09 07:13:22 it's correct partial fix, but not a full fix for my use-case 2014-12-09 07:13:42 ok we keep it then 2014-12-09 07:14:14 ncopa: hi, openvswitch is now up and running - works perfect ;) any plans on 3.1 release? cryptsetup needs still manual install from edge until that alsa openvswitch :( 2014-12-09 07:14:54 xen_roger: i was hoping to tag v3.1 today 2014-12-09 07:15:15 ;) nice to hear 2014-12-09 07:15:17 you can use v3.1 git repos if you want 2014-12-09 07:15:30 3.1.0_rc5 should be pretty usable 2014-12-09 07:16:11 ncopa, other option is to revert the commit that broke things in first place 2014-12-09 07:18:12 fabled: do you know which commit that broke it? 2014-12-09 07:18:42 ncopa, yes 2014-12-09 07:18:50 well, with 99% likelyness 2014-12-09 08:12:26 poor arm builder 2014-12-09 08:12:29 it will be busy today 2014-12-09 08:31:57 where can I find alpine for arm? 2014-12-09 08:43:06 got iy 2014-12-09 08:43:27 it, even 2014-12-09 08:43:36 :) 2014-12-09 08:44:33 heh, arm builder is still fighting with samba build 2014-12-09 08:44:50 it has no idea that when finally done, 2 kernels are waiting... 2014-12-09 08:45:12 rnalrd: could you please bump linux-virt-grsec too? 2014-12-09 08:45:18 sure 2014-12-09 08:45:22 thanks! 2014-12-09 08:45:35 go is actually not that long to build 2014-12-09 08:45:44 samba takes quite a while 2014-12-09 08:46:20 kernel was also not insane time; 30-60mins or so, iirc 2014-12-09 08:49:42 downloading armhf rc5 now. should I wait for 3.1 to test it? 2014-12-09 08:58:19 ScrumpyJack, rc5 works for me just fine 2014-12-09 09:10:06 how do you install it to sd card fabled ? 2014-12-09 09:10:27 normally i'd dd an image 2014-12-09 09:10:28 make regular vfat card, and just extract the tarball to it 2014-12-09 09:10:40 :( so it's all vfat? 2014-12-09 09:10:44 yes 2014-12-09 09:10:47 it's standard alpine 2014-12-09 09:10:49 it runs from tmpfs 2014-12-09 09:10:57 which is actually *good* 2014-12-09 09:11:13 the biggest problem on my rpi:s is that the sd-card wears out due to everyone using mmc/sd as root 2014-12-09 09:11:23 probably not too big issue on recent cards 2014-12-09 09:11:33 but i prefer having fully read-only access to mmc/sd 2014-12-09 09:12:17 so i'll need to use lbu? 2014-12-09 09:12:26 to write to the sd card? 2014-12-09 09:13:05 yes 2014-12-09 09:13:11 unless you setup-disk it 2014-12-09 09:22:13 ncopa, would it ok to update zeromq, https://raw.githubusercontent.com/zeromq/zeromq4-1/master/NEWS ? 2014-12-09 09:24:19 we have 4.0.5 already? 2014-12-09 09:24:24 4.1 2014-12-09 09:24:40 there are no stable 4.1 release? 2014-12-09 09:25:02 ok 2014-12-09 09:25:15 fabled: vfat on my SD, copied contents of rc5, marked as boot. nothing :( 2014-12-09 09:25:15 is it ABI compatible with 4.0? 2014-12-09 09:25:59 not checked, but seeing NEWS file seems new functions were added 2014-12-09 09:26:16 but it says "Release 4.1.0 is backwards-compatible with stable releases back to 2.2" 2014-12-09 09:26:46 ScrumpyJack, do you get the rpi rainbow splash screen? 2014-12-09 09:26:47 anyway, its rc1 2014-12-09 09:26:54 nope 2014-12-09 09:27:38 it behaves as if i don't have a card inserted 2014-12-09 09:27:38 so it's probably reading the config, but somehow failing to boot 2014-12-09 09:27:48 oh 2014-12-09 09:28:01 strange 2014-12-09 09:28:08 bad card? 2014-12-09 09:28:12 rpi is picky on them 2014-12-09 09:28:27 i do disable the rainbow screen on config.txt, though 2014-12-09 09:30:49 it is partitioned sd too, right? 2014-12-09 09:31:29 i think rpi wants sd that has traditional partition table. and it looks for vfat on the first partition 2014-12-09 09:31:44 what do you mean by partitioned sd? 2014-12-09 09:32:10 sorry 2014-12-09 09:32:16 fdisk 2014-12-09 09:32:29 or parted 2014-12-09 09:32:34 fdisk 2014-12-09 09:34:10 dos compat flag needed? 2014-12-09 09:35:32 duh! linux partition. which partition type should I use? 2014-12-09 09:38:10 ScrumpyJack, standard vfat 2014-12-09 09:38:48 so 0xb, 0xc, or 0xe depending on card size and what fat cluster size was used 2014-12-09 09:38:56 try 0xb first 2014-12-09 09:41:16 hmm, what's that in fdisk terms? 2014-12-09 09:41:31 I don't have a vfat partition type in fdisk 2014-12-09 09:41:38 0xb 2014-12-09 09:41:41 (I only ever use 83_ 2014-12-09 09:42:02 are we talking about the same thing? 2014-12-09 09:42:21 yes 2014-12-09 09:42:35 you can use '0xb' or just 'b' 2014-12-09 09:42:47 command 'l' lists them 2014-12-09 09:42:59 it's explained as "W95 FAT32" 2014-12-09 09:43:29 aha! W95 FAT32 = vfat? right 2014-12-09 09:43:31 /join virtualbox 2014-12-09 09:43:44 sry, missed 2014-12-09 09:44:35 ScrumpyJack, yes. technically vfat == fat with additional way to encode long filenames 2014-12-09 09:44:40 when using 0xb, i got empty space :) 2014-12-09 09:45:03 sorry, i know little about fat 2014-12-09 09:45:19 (but that i remember now) 2014-12-09 09:46:49 i need help with writing a wiki page about localization on Alpine Linux 2014-12-09 09:46:54 #598 2014-12-09 09:47:01 http://bugs.alpinelinux.org/issues/598 2014-12-09 09:47:46 basically, it can say: we only support utf-8 2014-12-09 09:47:56 and have a link to the timezone doc 2014-12-09 09:48:01 or similar 2014-12-09 09:48:40 ok, I don't mind looking into it after I do the logos for you, but i'm slow :) 2014-12-09 09:49:06 fabled: success! 2014-12-09 09:49:19 ScrumpyJack, nice! 2014-12-09 09:49:34 i made a no-interaction video player out of my rpi 2014-12-09 09:49:46 shows splash during boot, and starts omxplayer on startup 2014-12-09 09:49:55 quite pro looking boot 2014-12-09 09:50:54 fabled: a few errors going through openrc 2014-12-09 09:51:11 yes, one about rtc 2014-12-09 09:51:16 since the box does not have a rtc 2014-12-09 09:51:25 and other maybe on /lib/modules ? 2014-12-09 09:51:47 modprobe can't change directory to /lib/modules 2014-12-09 09:51:48 yes 2014-12-09 09:52:42 yeah, not fussed about the clock errors ;) 2014-12-09 09:53:17 i'll play later - back to work 2014-12-09 13:12:46 ERROR: dnscrypt-proxy: Failed to build 2014-12-09 13:12:52 checking 2014-12-09 13:13:14 it fails on libltdl 2014-12-09 13:13:16 umh 2014-12-09 13:17:06 bah, i've channged only the pre-install script 2014-12-09 13:19:02 ncopa, this is the problem: $LIBTOOLIZE --ltdl 2014-12-09 13:29:44 actually he doesn't find libltdl dir 2014-12-09 13:34:03 o kgot it 2014-12-09 13:34:06 ok got it 2014-12-09 15:23:05 please git pull git://git.alpinelinux.org/fab/aports 2014-12-09 15:23:12 there are some updates 2014-12-09 15:28:51 thanks 2014-12-09 22:26:43 someone can give any clue? http://hastebin.com/misejuqaje.lua 2014-12-09 23:20:54 I keep wondering why hastebin uses extensions like that 2014-12-09 23:22:48 alacerda: missing header, but I'm unsure which one 2014-12-09 23:22:54 obviously X11 related 2014-12-09 23:23:18 X11/extensions/XInput.h maybe 2014-12-09 23:25:05 good night 2014-12-10 09:46:50 hi barthalion sorry I was afk 2014-12-10 09:46:54 but thanks... 2014-12-10 09:47:39 I am search wich package i should install in order to have this header... tried some x11*-dev but nothing yet 2014-12-10 09:52:21 alacerda: try freeglut-dev 2014-12-10 10:08:01 k0r10n, worked... I got another error but it is another topic :) thnaks 2014-12-10 10:24:45 please hold your git pushes for a sec 2014-12-10 10:29:57 ncopa: yeahhh release time ;) 2014-12-10 10:30:11 Yay! 2014-12-10 10:31:05 it would be nice to have an rss feed to announce stuff like that ... no? yes? 2014-12-10 10:38:09 hum 2014-12-10 10:39:09 xen_roger: I think there is a fearture request for that on b.a.o 2014-12-10 10:39:52 \o/ 2014-12-10 10:39:59 http://bugs.alpinelinux.org/issues/1831 2014-12-10 10:41:01 I see ... +1 for point 2 of request ;) 2014-12-10 10:44:35 ncopa, 2014-12-10 11:05:07 does this look ok? http://wwwtest.alpinelinux.org/posts/Alpine-3.1.0-released.html 2014-12-10 11:05:54 what failed on armhf? 2014-12-10 11:06:02 build-3-1-armhf: failed to build v3.1.0_rc5: http://bld1.alpinelinux.org/buildlogs/build-3-1-armhf.v3.1.0_rc5.log 2014-12-10 11:06:10 why it even tried to build that old tag? 2014-12-10 11:06:13 i stopped it 2014-12-10 11:06:20 i messed upt i think 2014-12-10 11:06:34 algitbot> build-3-1-armhf: release v3.1.0 uploaded 2014-12-10 11:06:40 but should be released now 2014-12-10 11:08:08 ok 2014-12-10 11:08:12 thx 2014-12-10 11:15:45 This is the release s built with musl libc 2014-12-10 11:16:00 some typos 2014-12-10 11:16:24 This release is built with musl libc 2014-12-10 11:16:49 with v2.x versions 2014-12-10 11:17:22 inside my head i raed "with version two point x versions" 2014-12-10 11:17:34 best to remove "versions" 2014-12-10 11:18:08 for information *on* how to perform 2014-12-10 11:18:58 available for v3.1 => available *in* v3.1 2014-12-10 11:21:01 otherwise, looks good :) 2014-12-10 11:21:10 wanna mantion kde? 2014-12-10 11:22:35 also, perhaps mention some more non desktop applications in new features 2014-12-10 11:28:10 ah, kde is in edge testing :) 2014-12-10 11:34:59 only 50 commits, I was after ncopa last time :D 2014-12-10 12:10:10 typos fixed: http://wwwtest.alpinelinux.org/posts/Alpine-3.1.0-released.html 2014-12-10 12:10:12 ok now? 2014-12-10 12:22:48 This *release* is built" 2014-12-10 12:28:04 just booted alpine-xen 3.1 2014-12-10 12:28:17 not a single error :) brilliant! 2014-12-10 12:34:27 ncopa, when are we going to release 3.1? Is the date already set? 2014-12-10 12:34:44 fcolista, today. images are ready already 2014-12-10 12:35:10 amazing 2014-12-10 12:35:33 were 1 month late :-( 2014-12-10 12:36:00 I didn't hear nobody knocking on the door ;) 2014-12-10 14:08:35 hi! just wanted t report a few typos/mistakes on one of the wiki pages 2014-12-10 14:09:25 http://wiki.alpinelinux.org/wiki/FAQ#Windows_clients_reports_an_error_when_trying_to_sync 2014-12-10 14:09:58 1) should this perhaps be "client reports" or "clients report"? 2014-12-10 14:11:06 2) "before it is satisfied it is in sync" -- seems like a typo or editing residue... 2014-12-10 14:11:47 3) "15-20mins" --> a space before 'mins'? 2014-12-10 14:11:56 I guess that's all for now... 2014-12-10 14:17:54 i'm looking at binutils (use ld.gold by default instead of ld.bfd), gcc 4.9.2, and few other major updates for edge soonish (maybe tomorrow / the day after) 2014-12-10 14:18:01 building currently gcc-4.9.2 2014-12-10 14:18:55 needed to refresh / rebase a whole lot of patches 2014-12-10 14:18:59 but seems it built now. 2014-12-10 14:19:08 next step is to see if it still works :D 2014-12-10 14:36:01 is util-linux on your list? 2014-12-10 14:36:28 no 2014-12-10 14:36:33 just core toolchain 2014-12-10 14:36:36 k 2014-12-10 14:36:53 binutils + gcc + musl + libc-dev (splitting up bsd headers to separate bsd-compat-headers) 2014-12-10 14:48:28 ok, binutils+gcc+musl are pretty much ready for push 2014-12-10 14:48:36 i'll do still libc-dev split to bsd-compat-headers 2014-12-10 14:48:41 and push tomorrow 2014-12-10 14:48:48 it's almost time for me to go 2014-12-10 14:59:07 hum 2014-12-10 14:59:12 i ont konw what happenend 2014-12-10 15:00:18 why they fail on _rc5 tag? 2014-12-10 15:00:43 somethign is wrong... 2014-12-10 15:02:04 oh no! what? 2014-12-10 15:02:11 builders 2014-12-10 15:02:37 i don't trust builders, they mess up my kitchen extention ;) 2014-12-10 15:03:16 you mean this? build/build-3-1-x86 (null) 2014-12-10 15:03:33 ok, binutils made to default to gold 2014-12-10 15:03:52 it was able to link gcc 4.9.2, so assume i got config right 2014-12-10 15:04:01 i'm waiting until tomorrow for gcc update 2014-12-10 15:04:02 i think 2014-12-10 15:04:11 or maybe i push it late today 2014-12-10 15:04:24 i also need to figure out libc-dev / bsd-compat-headers split 2014-12-10 15:04:30 i'd like to do that before gcc update 2014-12-10 15:04:47 new binutils is a lot bigger 2014-12-10 15:04:54 due to having both: gold + bfd linker 2014-12-10 15:05:56 the v3.1.0 tag is not visible from 3.1-stable branch 2014-12-10 15:05:59 :-( 2014-12-10 15:06:23 i think i branched it before I tagged it 2014-12-10 15:06:28 which was a mistake 2014-12-10 15:15:38 still the annoying lvm error messages :( 2014-12-10 15:39:30 ncopa: any idea about that error? 2014-12-10 15:40:53 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0curl: (56) Recv failure: Connection reset by peer 2014-12-10 15:41:04 it could not fetch sources 2014-12-10 15:41:09 probably a temp issue 2014-12-10 16:00:34 ncopa: im trying 3.1 gst python, but get ImportError: Error relocating /usr/lib/python2.7/site-packages/gst-0.10/gst/_gst.so: libxml_xmlDocPtrWrap: symbol not found 2014-12-10 16:02:39 ugh 2014-12-10 16:03:03 libxml got updated? 2014-12-10 16:03:11 i'm tempted to push gcc 4.9.2 on edge 2014-12-10 16:03:23 could also be underlinking issue 2014-12-10 16:03:32 fabled: i'm ok with pushing it 2014-12-10 16:03:42 gcc-4.9.2 sounds cool :) 2014-12-10 16:24:21 here we go 2014-12-10 16:24:25 let's see what blows up 2014-12-10 16:24:31 though, i need to run 2014-12-10 16:24:54 hope it builds on arm too 2014-12-10 16:25:03 x86_64 i tested several times, but not others so well 2014-12-10 16:25:06 bye 2014-12-10 16:41:30 hah 2014-12-10 17:31:31 ncopa, or others: how can I create an "bootable" rootfs kernel no needed ... like debootstap on debian, but for alpine ? 2014-12-10 17:40:18 nobody? 2014-12-10 17:49:54 xen_roger: I use linux containers to do that 2014-12-10 17:50:34 what is your host OS? 2014-12-10 17:50:40 mhh... need an armhf rootfs 2014-12-10 17:50:52 host-os could be everything 2014-12-10 17:51:04 not gonna happen :) 2014-12-10 17:51:21 what's your host OS architecture? 2014-12-10 17:51:45 x86 2014-12-10 17:52:52 also I have a armhf running debian ... if native arch is needed ... 2014-12-10 17:52:56 any idea? 2014-12-10 17:54:47 so you have an armhf chip or you emulating? 2014-12-10 17:55:01 have arm 2014-12-10 17:56:15 in that case apt-get install lxc 2014-12-10 17:56:41 then lxc-create -n alpine -t path/to/template 2014-12-10 17:58:06 template? where to getß 2014-12-10 17:58:10 LXC runs very well - I have a container for x86 & another for amd64 2014-12-10 17:58:41 just checked debian and templates are in apt-get install lxc-templates 2014-12-10 17:59:04 same in alpine 2014-12-10 17:59:04 then lxc-start -n alpine 2014-12-10 17:59:34 you'll end up at the alpine prompt 2014-12-10 17:59:50 log in a root, no password :) 2014-12-10 18:00:02 type halt 2014-12-10 18:00:11 don't get the point ... means the template is the problem? is there one in lxc-templates? 2014-12-10 18:00:19 yes 2014-12-10 18:00:33 lots in lxc-templates 2014-12-10 18:03:08 they are in /usr/share/lxc/templates/ 2014-12-10 18:03:30 I doubt that they work for alpine arm as expected 2014-12-10 18:03:45 barthalion: this is a great way to test :) 2014-12-10 18:03:58 feel free to quote me later 2014-12-10 18:05:48 http://wiki.alpinelinux.org/wiki/LXC 2014-12-10 18:38:26 thanks so far at all ... 2014-12-10 18:54:10 xen_roger: did it work? 2014-12-10 19:42:39 ncopa, seems updating gcc is problematic. 2014-12-10 19:42:52 apk upgrade is run only after each commit is fully processed 2014-12-10 19:43:15 but during gcc bump, gcc should be rebuild, apk upgrade run, and java-gcj-compat rebuild 2014-12-10 19:43:25 java-gcj-compat gets it's version automatically from gcc 2014-12-10 19:45:16 i wonder why x86 worked 2014-12-10 19:45:20 i did fix x86_64 by hand 2014-12-10 19:45:26 armhf is still rebuilding gcc 2014-12-10 19:46:53 good night 2014-12-10 19:58:46 ScrumpyJack: after a lot of trouble ... rootfs is created and "starts up" - BUT - it ist totally buggy... 2014-12-10 20:04:55 why does control+c rebooting the system ? 2014-12-10 20:08:38 no idea. i have a pi at work i can try 2014-12-10 20:10:31 it would be nice to run alpine on arm ... but the way to create rootfs is not that easy :-/ and now it is unuseable :( 2014-12-10 20:26:26 you could get a rpi and download http://wiki.alpinelinux.org/cgi-bin/dl.cgi/v3.1/releases/armhf/alpine-rpi-3.1.0-armhf.rpi.tar.gz 2014-12-11 04:47:06 ncopa: mksite, reduce css validators error patch, http://sprunge.us/FcbJ 2014-12-11 05:15:21 http://sprunge.us/chhd <- minor layout improvent for mobiles 2014-12-11 05:16:06 pls, let me know if needed to send via ML 2014-12-11 05:24:05 can someone tell me "when and how" in style.css header svg gets used ? 2014-12-11 05:45:33 pls ignore the chhd patch, apply http://sprunge.us/KSAe instead 2014-12-11 07:33:06 I got this error: 2014-12-11 07:33:12 /usr/lib/gcc/x86_64-alpine-linux-musl/4.8.3/../../../../x86_64-alpine-linux-musl/bin/ld: cannot find -lSM 2014-12-11 07:33:43 in other Alpine Machine I can "make" with no problem 2014-12-11 07:33:51 alacerda: libsm? 2014-12-11 07:34:03 there I have the SM 2014-12-11 07:34:06 http://hastebin.com/uhiquroceq.hs 2014-12-11 07:34:07 oops 2014-12-11 07:34:16 AmatCoder, let me try 2014-12-11 07:35:22 AmatCoder, I am testing... but in this mean time, let me ask you sth 2014-12-11 07:35:33 on this hastebin link that I sent you... 2014-12-11 07:35:40 did you see the line: 2014-12-11 07:35:42 apk info --who-owns /usr/include/X11/SM 2014-12-11 07:36:07 the apk answer was: 2014-12-11 07:36:07 ERROR: /usr/include/X11/SM: Could not find owner package 2014-12-11 07:36:22 hum... shouldn't it answer to me: libsm ? 2014-12-11 07:36:34 hmm, it does not work directories currently 2014-12-11 07:36:44 the reason is that one directory can have "multiple" owners 2014-12-11 07:37:57 fabled, gotcha 2014-12-11 07:39:25 will this behavior last forever or... I mean, wouldn't it be "better" if it lists all these multiple owners? 2014-12-11 07:39:57 yes 2014-12-11 07:40:04 i remember looking at that 2014-12-11 07:40:46 i think the problem was that implementing that required extending the internal directory database structures, or enumerating the whole directory tree to find directory entry 2014-12-11 07:41:24 but yes, it should be still fixed 2014-12-11 07:41:49 at least an error like this: "ERROR: /usr/include/X11/SM: Is a directory" 2014-12-11 07:41:51 and would be nice to show all the permissions each package wanted for the dir, and what was the conclusion 2014-12-11 07:43:47 ah ok. cool 2014-12-11 07:43:48 ;) 2014-12-11 07:49:54 AmatCoder, thanks libsm-dev owrked 2014-12-11 07:50:09 nice 2014-12-11 08:05:10 kunkku: http://sprunge.us/DLCE 2014-12-11 08:09:32 ncopa, was the mksite patch useful ? 2014-12-11 08:10:27 I think I may have one more 2014-12-11 08:11:48 vkrishn: i'll look at it soonish 2014-12-11 08:12:51 ok 2014-12-11 08:12:57 how about name "fractal" for your alpine-site generation software ? 2014-12-11 08:13:24 for alpine-mksite 2014-12-11 08:23:25 I have signed the releases: http://nl.alpinelinux.org/alpine/v3.1/releases/x86_64/ 2014-12-11 08:23:32 there are a .sig 2014-12-11 08:33:44 nice 2014-12-11 08:33:55 did you publish your key / fingerprint somewhere? 2014-12-11 08:34:20 http://pgp.mit.edu/pks/lookup?search=ncopa&op=index 2014-12-11 08:34:29 07D9495A 2014-12-11 08:34:34 you need to publish it on www.alpinelinux.org too 2014-12-11 08:34:39 well, it'd be nice to 2014-12-11 08:34:49 i can sign your key meanwhile 2014-12-11 08:37:52 ncopa: could you spend me a few minutes ? 2014-12-11 08:38:07 xen_roger: sure 2014-12-11 08:38:16 ncopa: what do you use for building arm pkgs? 2014-12-11 08:38:38 wandboard quad 2014-12-11 08:38:42 still trying to setup armhf-rootfs but no luck so far :( 2014-12-11 08:39:19 i actualy dont have arm board physically here 2014-12-11 08:39:24 ...yet 2014-12-11 08:39:53 xen_roger: what hw are you using? 2014-12-11 08:40:20 fabled: maye should publish the apk keys too on alpinelinux.org 2014-12-11 08:40:23 rk3066 based arm 2014-12-11 08:40:29 at least make them available 2014-12-11 08:40:54 xen_roger, what's the issue? 2014-12-11 08:41:07 we don't ship kernel for that though 2014-12-11 08:41:11 but the userland should work 2014-12-11 08:41:37 ok: host is x86 and arm is cortex-a9 ... bare ... no os on it ... where to start 2014-12-11 08:42:08 debian I could use debootstrap and qemu-static ... what is the alpine way to setup an rootfs? 2014-12-11 08:42:24 we don't have qemu-static support 2014-12-11 08:42:31 well 2014-12-11 08:42:42 it could probably be made work with some minor patching 2014-12-11 08:43:11 how is the pi-rootfs created? 2014-12-11 08:43:24 it's built natively on wandboard arm quadcore 2014-12-11 08:43:45 ok, and how? 2014-12-11 08:43:47 i have one board like that; ncopa is getting few more 2014-12-11 08:44:05 what exactly you mean with 'how' ? 2014-12-11 08:44:19 the rpi image building is part of alpine-iso scripts 2014-12-11 08:44:42 i have full alpine install on my wandboard; it's hand built, as we don't have images for wandboard yet 2014-12-11 08:45:05 what comes to rk3066 2014-12-11 08:45:18 we'd need to enable support for it in linux-grsec, and u-boot 2014-12-11 08:45:42 after that you should be able to run alpine on it; though making a boot media would probably need to be done by hand 2014-12-11 08:46:16 i'm really hoping we get images for arm+u-boot 2014-12-11 08:46:20 ok, the kernel is build and running ... it starts into init from sd-card ... so the only thing I need is the rootfs 2014-12-11 08:46:57 for experimenting 2014-12-11 08:47:03 like emerge stage3 on gentoo or debootstrap on debian ... 2014-12-11 08:47:11 morning 2014-12-11 08:47:17 hi ScrumpyJack 2014-12-11 08:47:34 you could on x86/x86_64 do something like: apk --root arm-root --arch armhf --initdb --no-scripts add alpine-base 2014-12-11 08:47:35 ScrumpyJack: still on arm-rootfs building :( 2014-12-11 08:48:04 but for rpi we ship standard alpine boot media: that is, the run from tmpfs 2014-12-11 08:48:18 and that's how my wandboard works too 2014-12-11 08:48:35 it's tmpfs host, running builders in lxc containers 2014-12-11 08:49:07 ok ... let me try the one menton above ... is there a docu somewhere ? about "apk --root arm-root --arch armhf --initdb --no-scripts add alpine-base " 2014-12-11 08:49:11 no 2014-12-11 08:49:18 :-/ 2014-12-11 08:49:41 you're welcome to write a wiki page about that ;) 2014-12-11 08:49:49 we just barely got the rpi image building done in time 2014-12-11 08:49:56 well.. we were late on that 2014-12-11 08:50:13 ok ... when I am done ... and it works I will do 2014-12-11 08:50:51 got 1 month delayed, mostly due to arm :-/ 2014-12-11 08:51:10 lesson learned: start up the arm build servers first 2014-12-11 08:51:12 yeah. waiting for the slow builders, and doing the image creating 2014-12-11 08:51:19 and do it early 2014-12-11 08:51:48 i think i will start the arm servers first or second week in april 2014-12-11 08:52:22 meaning that toolchain freeze is 1 april 2014-12-11 08:54:19 well, i think we got the big toolchain changes already in edge: gcc-4.9.2, binutils/gold-as-default 2014-12-11 08:54:26 though arm is still rebuilding gcc 2014-12-11 08:55:07 it's on stage3 now 2014-12-11 08:56:01 ncopa, for dmvpn fix: http://patchwork.ozlabs.org/patch/419970/ 2014-12-11 08:57:11 fabled: i push that now or we wait for upstream feedback? 2014-12-11 08:57:22 we could wait for some feedback 2014-12-11 08:57:27 fabled: apk --root my-rootfs/ --arch armhf --initdb --no-scripts add alpine-base << got error ERROR: unsatisfiable constraints: alpine-base (missing): required by: world[alpine-base] 2014-12-11 08:57:49 xen_roger, mmm, you probably need also "--repository -U" 2014-12-11 08:57:55 add --repositories --keys-dir 2014-12-11 08:58:03 oh, keys-dir too 2014-12-11 08:58:20 *crying* 2014-12-11 08:58:37 --repositories-file or --reporitory 2014-12-11 08:58:48 there's http://wiki.alpinelinux.org/wiki/Installing_Alpine_Linux_in_a_chroot 2014-12-11 08:58:55 giving an idea onhow to do chroots 2014-12-11 08:59:05 you'll just need the --arch armhf and --no-scripts in addition 2014-12-11 08:59:33 beenthinking of an albootstrap script 2014-12-11 08:59:46 which is wrapper to apk add --initdb 2014-12-11 09:01:07 << will try 2014-12-11 09:08:16 how do we feel about not announcing armhf to the rpi community just yet? 2014-12-11 09:09:52 last night i started looking at where to announce, collecting contact details and putting together a "press pack" but is it "ready"? 2014-12-11 09:10:14 first thing i did this morning on a fresh rpi install failed :) 2014-12-11 09:10:27 Doesn't get it working ... could anyone provide the commandline I need? maybe I did something wrong :/ 2014-12-11 09:11:08 currently: apk --root my-rootfs --repository http://dl-5.alpinelinux.org/alpine/v3.1/main --arch armhf --initdb --no-scripts add alpine-base => failed 2014-12-11 09:11:51 ncopa: pushed 2014-12-11 09:12:59 kunkku: thanks 2014-12-11 09:13:25 xen_roger: it failed here too while setting permissions. do you need run it as non-root? 2014-12-11 09:14:17 ncopa: yes, running as root get: ERROR: unsatisfiable constraints: alpine-base (missing): required by: world[alpine-base] 2014-12-11 09:14:58 you need --update to have it force download the repo cache 2014-12-11 09:15:05 but it will fails on the keys too 2014-12-11 09:15:50 this worked here: apk add alpine-base --root my-rootfs/ --arch armhf --initdb --no-scripts --repository http://dl-5.alpinelinux.org/alpine/v3.1/main --keys-dir /etc/apk/keys --update 2014-12-11 09:16:15 :-* mercy ! 2014-12-11 09:16:26 if you dont have the keys dir you can use --allow-untrusted 2014-12-11 09:16:36 works... 2014-12-11 09:17:30 i should write an albootstrap wrapper script for that 2014-12-11 09:17:59 would be nice ... 2014-12-11 09:18:08 which can download the static apk if needed too, like kunkku did for the alpine lxc template script 2014-12-11 09:23:47 ncopa, re: libc-dev, bsd-compat-headers split 2014-12-11 09:23:55 do we keep libc-dev still? 2014-12-11 09:24:05 it's sole purpose is to provide dependency for musl-dev 2014-12-11 09:24:13 i guess that's still the way to do it 2014-12-11 09:24:55 that change will break source builds for various packages 2014-12-11 09:25:01 should probably mention that on alpine-list 2014-12-11 09:26:46 should i add that sys/tree.h or what was it 2014-12-11 09:28:38 i'll update sys/queue.h too 2014-12-11 09:34:33 yeah, it will break source builds for various 2014-12-11 09:34:41 we will have to fix them when they show up 2014-12-11 09:34:44 ncopa, http://sprunge.us/WHGj 2014-12-11 09:41:31 hum 2014-12-11 09:41:53 the subpackage is only added if CLIBC=musl 2014-12-11 09:42:05 yes. glibc ships it's own, it'd conflict 2014-12-11 09:42:17 or you think it should be unconditional? 2014-12-11 09:42:23 hmm 2014-12-11 09:42:35 i think dhcpcd will need it 2014-12-11 09:42:40 perhaps we should always create the pkg, and include the files in musl 2014-12-11 09:43:20 i dont want if musl then add bsd-compat-headers to make depends 2014-12-11 09:43:32 true 2014-12-11 09:43:44 how about letting it be empty for non-musl? 2014-12-11 09:44:17 case "$CLIBC" in 2014-12-11 09:44:17 musl) 2014-12-11 09:44:19 ncopa, http://sprunge.us/AiML 2014-12-11 09:44:23 install ... 2014-12-11 09:44:30 yeah 2014-12-11 09:44:45 that will work 2014-12-11 09:45:17 and the preference on how to fix packages is: 2014-12-11 09:45:23 1) patch to not requires bsd headers 2014-12-11 09:45:30 2) if too complicated, depend on bsd-compat-headers 2014-12-11 09:45:37 yup 2014-12-11 09:46:24 i think i'll just add bsd-compat-headers to makedepends 2014-12-11 09:46:31 since it will likely require refactoring 2014-12-11 09:46:36 ? 2014-12-11 09:46:41 sys queue 2014-12-11 09:46:53 ah right 2014-12-11 09:46:54 yes 2014-12-11 09:46:55 will require that code is refactored 2014-12-11 09:47:17 i think you can test it on dhcpcd 2014-12-11 09:47:21 maybe openrc too 2014-12-11 09:47:42 asterisk too, there's security update for it 2014-12-11 09:47:57 quagga should have configure magic, and it should work without bsd-compat-headers 2014-12-11 09:48:05 though it uses sys/cdefs.h if it exists 2014-12-11 09:56:05 hmmh 2014-12-11 09:57:06 ada still failed 2014-12-11 10:02:26 >>> qt5-qtdeclarative: Unpacking /var/cache/distfiles/qtdeclarative-opensource-src-5.3.2.tar.xz... 2014-12-11 10:02:27 Error relocating /usr/lib/libstdc++.so.6: : symbol not found 2014-12-11 10:03:05 hum 2014-12-11 10:03:45 i wonder if it is related #3616 2014-12-11 10:04:16 check if that's edge or 3.1-stable 2014-12-11 10:04:24 could be new gcc shipping different libstrc++ 2014-12-11 10:04:28 food 2014-12-11 10:18:42 fabled, re #3607 http://sprunge.us/XBFe 2014-12-11 10:19:07 but it becomes 5Mb 2014-12-11 10:19:16 is currently 600-700k 2014-12-11 10:21:25 anybody tried https://github.com/micropython/micropython if useful 2014-12-11 10:25:31 we have a couple boards, and i prefer python a lot, but on arm32 i like to stick with my gcc and some handcrafted asm. but i guess for beginners on embedded microcontrollers this is very very convenient. 2014-12-11 10:43:33 ncopa, java works on my edge/x86_64 2014-12-11 10:43:46 with new gcc 2014-12-11 10:45:54 works on mine to 2014-12-11 10:46:00 i wonder if its x86 only 2014-12-11 11:05:51 x86 fails on qt5-qtdeclarative 2014-12-11 11:06:17 Error relocating /usr/lib/libstdc++.so.6: : symbol not found 2014-12-11 11:06:18 hum 2014-12-11 11:06:24 sounds like x86 issue 2014-12-11 11:06:28 same as java 2014-12-11 11:08:16 yes 2014-12-11 11:11:06 seems stdc++ is completely broken 2014-12-11 11:11:11 even hello world does notwork 2014-12-11 11:11:53 weird 2014-12-11 11:11:56 there's no symbol name 2014-12-11 11:11:59 Error relocating /usr/lib/libstdc++.so.6: : symbol not found 2014-12-11 11:12:06 as if the symbol is "" or empty string 2014-12-11 11:12:39 hm i've seen that once.. i think it was some linker weirdness 2014-12-11 11:21:43 hmm 2014-12-11 11:21:49 gold seems to generate symversions 2014-12-11 11:21:58 but i compiled gcc with disabled symversions 2014-12-11 11:23:09 though things work on x86_64 2014-12-11 11:23:54 Symbol table '.dynsym' contains 4207 entries: 2014-12-11 11:23:54 Num: Value Size Type Bind Vis Ndx Name 2014-12-11 11:23:54 0: 00000000 0 NOTYPE LOCAL DEFAULT UND 2014-12-11 11:24:08 1: 000cfe80 0 SECTION LOCAL DEFAULT 17 2014-12-11 11:24:13 2: 0005039c 0 NOTYPE GLOBAL DEFAULT 9 _init 2014-12-11 11:24:13 3: 0009fb87 0 NOTYPE GLOBAL DEFAULT 12 _fini 2014-12-11 11:24:18 ... 2014-12-11 11:24:31 so yeah, seems that the first symbol is local, undefined and no string name 2014-12-11 11:43:37 ncopa seems https:// is still going to older webserver or old webfolder 2014-12-11 12:05:57 crow: yes 2014-12-11 12:06:13 so we still can access the old content 2014-12-11 12:16:48 ncopa on can you fix then the Download latest release 2014-12-11 12:17:00 for arm rpi i just used this and works fine: http://dpaste.com/08HGW7B 2014-12-11 12:24:12 is there a deadline for AL3.2? 2014-12-11 12:24:42 (i'd like to finish some things in time for 3.2) 2014-12-11 12:39:35 hmmm... got gcc internal compiler error on arm now :-/ 2014-12-11 12:57:40 ScrumpyJack: yes. toolchain (and build tools) freeze is 1 april 2014-12-11 12:57:46 together with ABI freeze 2014-12-11 12:57:58 feature freeze is 15 april 2014-12-11 12:58:33 and the lets-shoot-the-engineers-and-just-ship-the-product deadline is 1 May 2014-12-11 13:06:29 ok cool. 2014-12-11 13:22:08 poor engineers 2014-12-11 13:26:12 speaking of new gcc, assuming it will work 2014-12-11 13:26:29 which -fstack-protector are we using now? 2014-12-11 13:36:18 barthalion, -fstack-protector-strong 2014-12-11 13:36:48 ncopa pign 2014-12-11 13:37:06 ncopa i know you've left uclibc in the dust, but did you ever get anywhere with this -> http://lists.uclibc.org/pipermail/uclibc/2012-October/047059.html 2014-12-11 13:37:35 i hit this issues and traced it back to do_dlclose and found that you hit the same thing 2014-12-11 13:42:51 blue|chrome: is it vlc? 2014-12-11 13:43:15 ncopa no i hit it while building librsvg 2014-12-11 13:43:24 but it is the same bug 2014-12-11 13:43:30 reverting the commit fixes i 2014-12-11 13:43:32 it 2014-12-11 13:43:55 which commit? 2014-12-11 13:44:01 reverting this commit here -> http://git.uclibc.org/uClibc/commit/ldso?h=0.9.33&id=9b42da7d0558884e2a3cc9a8674ccfc752369610 2014-12-11 13:45:45 more specifically it was while running gdk-pixbuf-loaders on libpixbufloader-svg.la 2014-12-11 13:46:05 ok 2014-12-11 13:46:12 which dlopens a bunch of libraries and it segfaults while trying to dlclose libbz2.so 2014-12-11 13:46:14 seem liks fabled fixed it for us 2014-12-11 13:46:17 we have a patch for it 2014-12-11 13:46:31 i'd like to see that 2014-12-11 13:46:59 http://git.alpinelinux.org/cgit/aports/commit/main/libc0.9.32/uclibc-dlclose-fix.patch?h=2.7-stable&id=d36e402fae2b31ca2bf6eafbafa77d716ea99b15 2014-12-11 13:47:00 it happens only on 64-bits so the unsigned int needs to be ElfW(Addr) 2014-12-11 13:47:31 http://git.alpinelinux.org/cgit/aports/tree/main/libc0.9.32/uclibc-dlclose-fix.patch?h=2.7-stable 2014-12-11 13:47:58 i think there are more issues in uclibc threading implementation 2014-12-11 13:48:07 which we concluded to be unfixable 2014-12-11 13:48:12 and switched to musl... 2014-12-11 13:49:10 ncopa i'm afraid you might be right because those patches are still misse a few other points i uncovered 2014-12-11 13:49:58 thanks for those links, i'll see if they help 2014-12-11 13:50:08 Is there something up with alpinelinux.org/packages/ ? I tried searching and get error 2014-12-11 13:50:32 Frosh: what error do you get? 2014-12-11 13:51:01 http://pastebin.com/JCHZB4kf 2014-12-11 13:52:10 :-D 2014-12-11 13:52:18 StatusText: Request Entity Too Large 2014-12-11 13:52:26 its an application error 2014-12-11 13:56:35 I can't type python? or both my chrome and firefox? 2014-12-11 13:57:26 got this compile error for v4l-utils: http://pastebin.com/wvk8PxhH. Re "int open64" (row 8) i can patch and set it as "open" (as it is defined in /usr/include/fnctl.h). But what about ioctl? 2014-12-11 13:58:18 sounds like it tries to override open and ioctl... 2014-12-11 13:58:20 hum 2014-12-11 13:58:39 right. Using a non std function 2014-12-11 13:58:45 that i salready defined. 2014-12-11 13:59:13 Tried a different computer to check 2014-12-11 13:59:15 i think that using "open" would work, but ioctl is defined with different args 2014-12-11 13:59:46 rather than int,int uses int, unsigned long int 2014-12-11 14:00:34 umh 2014-12-11 14:02:11 ncopa, if i patch LIBV4L_PUBLIC ioctl function calling it LIBV4L_ioctl , then i can grep this ioctl and set LIBV4L_ioctl 2014-12-11 14:02:27 would it work? 2014-12-11 14:02:59 umh 2014-12-11 14:03:01 too much 2014-12-11 14:03:07 maybe is better remove include 2014-12-11 14:23:59 hmm.. gcc issue on arm is likely: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63740 2014-12-11 14:24:04 retrying with the patch applied 2014-12-11 14:24:05 sigh 2014-12-11 14:41:12 vkrishn: i'd like to do this for small screen: http://responsivenavigation.net/examples/clean-grid/index.html 2014-12-11 15:03:36 I think I'd seen something similar, Griddle (necolas.github.io) 2014-12-11 15:03:48 what minimum browsers versions are we targetting ? 2014-12-11 15:13:47 or best just send me design png, with style guide, would convert it into html 2014-12-11 15:47:20 vkrishn: i'd like the home/downloads/about/community nav links be on a 2x2 grid when they don't fit on a single line 2014-12-11 15:48:28 ok, would give a try, using this css lib 2014-12-11 15:52:39 please not any css libs that are 100+kb 2014-12-11 15:56:43 this sites css is around 20k, and only part should be applying to the menu 2014-12-11 16:01:30 they do it with 12 lines css here: http://responsivenavigation.net/examples/clean-grid/index.html 2014-12-11 16:05:50 vkrishn: your changes are pushed to wwwtest.alpinelinux.org 2014-12-11 16:07:19 if you around for another 10-15mins I can send another patch 2014-12-11 16:09:00 or just tomorrow , with new 2x2 grid nav 2014-12-11 16:24:10 ok, I have 2x2 first cut, send patch ? 2014-12-11 16:25:51 just couple of lines 2014-12-11 16:42:52 replace styles.css with http://sprunge.us/KdSX 2014-12-11 16:50:22 or patch, https://dpaste.de/RbB4/raw 2014-12-11 17:00:38 @media all and (max-width:24.000em) 2014-12-11 17:00:41 should be better 2014-12-11 19:47:33 hey guys 2014-12-11 19:47:49 I had a few questions about your workflow when it comes to kernel development 2014-12-11 19:48:00 I'm new to this 2014-12-11 19:51:17 someone know how ulimit works on alpine/ 2014-12-11 19:51:18 ? 2014-12-11 23:58:32 ping ncopa 2014-12-12 06:43:28 finally. seems gcc-4.9.2 on arm compiled 2014-12-12 07:04:17 wasn't the gcc 4.9 critisized for for lots of regressrions ? 2014-12-12 07:04:43 regressions 2014-12-12 07:04:43 jarruda: pong 2014-12-12 07:04:46 morning 2014-12-12 07:05:22 vkrishn: yes it was. that is why we have waited til the .2 release 2014-12-12 07:05:37 ok 2014-12-12 07:07:25 ncopa, could pls drop this styles.css in wwwtest https://dpaste.de/4i38 2014-12-12 07:09:57 sporgj: what makes you think this is a good channel to ask about kernel development? 2014-12-12 07:12:39 http://www.alpinelinux.org/packages/ still gives me 413 error on a different computer 2014-12-12 07:13:05 only when I search for something 2014-12-12 07:13:08 same here 2014-12-12 07:14:25 fabled RPI pcm512x 1-004c: Failed to reset device: -5 , whole dmesg http://sprunge.us/aSVX 2014-12-12 07:14:41 fabled its on 3.1.0 2014-12-12 07:15:04 Frosh: Chromium? 2014-12-12 07:15:31 crow: ttps://github.com/raspberrypi/firmware/issues/273 2014-12-12 07:15:41 both firefox and chrome 2014-12-12 07:17:30 Well, Firefox works for me. 2014-12-12 07:20:58 ncopa, which one would be better, @media all and (max-width:48.000em) 2014-12-12 07:21:01 or @media all and (max-width: 400px) ? 2014-12-12 07:21:27 once ok, would cleanup and send as patch 2014-12-12 07:25:41 any border around nav links ? 2014-12-12 07:34:31 fabled thanks blacklisting the module does fix this msg in dmesg. would be better to add it to alpinelinux rpi iso as default? 2014-12-12 07:34:48 one more msg which i have is about: F2FS-fs (loop0): Magic Mismatch, valid(0xf2f52010) - read(0x31c79627) 2014-12-12 07:37:13 crow, seems that pcm512x is used on 'hifiberry' add-on board for rpi 2014-12-12 07:37:34 crow: I'm unsure if anyone tested f2fs 2014-12-12 07:38:08 barthalion, crow : i think that happens if modules=squashfs is not specified and it tries to load all file system drivers in turn 2014-12-12 07:38:19 vkrishn: i prefer em over px: http://stackoverflow.com/questions/609517/why-em-instead-of-px 2014-12-12 07:38:34 that's probably why you see the error about /lib/modules too 2014-12-12 07:38:49 does adding "modules=squashfs" boot param to cmdline.txt fix that? 2014-12-12 07:39:11 I mean 400px translated to em, or 720px ? 2014-12-12 07:39:20 barthalion this is an clean rpi image from alpinelinux, i did not changed anything thus i am reporting i :), 2014-12-12 07:39:30 fabled let me try it 2014-12-12 07:39:54 or it could be just that rpi config has f2fs enabled, but our regular config not 2014-12-12 07:39:55 ncopa: http://sprunge.us/dKNT let me know whether I can push it 2014-12-12 07:41:05 barthalion: you'll have to rebase it against the 'wwwtest' branch 2014-12-12 07:41:36 fabled seems i need to remount boot partition? '/media/mmcblk0p1/cmdline.txt' is read only 2014-12-12 07:41:51 yes, alpine keeps boot-media read-only 2014-12-12 07:41:59 ncopa: hm, ok, I guess some of these were fixed already 2014-12-12 07:42:05 that makes your boot-media long lived :) 2014-12-12 07:42:08 barthalion: we rewrote parts of it: http://wwwtest.alpinelinux.org/about 2014-12-12 07:42:10 mount -o remount,rw /media/mmcblk0p1 2014-12-12 07:42:56 ncopa, care to add 'sig' link to the front-page's download of default image too ? 2014-12-12 07:43:58 or other way to decide is, is the 2x2 needed on tablets ? 2014-12-12 07:44:05 fabled: sure, np. what should be in the hover text? 2014-12-12 07:44:16 coz, I think single row on tablets would be ok 2014-12-12 07:44:34 fabled and to remount back to read-only : mount -o remount,ro /media/mmcblk0p1 ? 2014-12-12 07:44:46 then, @media all and (max-width:32.000em) , should be ok 2014-12-12 07:45:13 ncopa, the file contents? 2014-12-12 07:45:15 oh 2014-12-12 07:45:21 it's binary signature 2014-12-12 07:45:24 not ascii armored 2014-12-12 07:45:46 should I sign with -a instead? 2014-12-12 07:45:57 fcolista, would you able to push xapian to main/ ? 2014-12-12 07:46:05 i think they are often ascii armored. it's not big dealt hough 2014-12-12 07:46:27 .asc as suffix? 2014-12-12 07:46:28 sometimes people post the ascii armor signatures to mailing list 2014-12-12 07:46:38 yes, .asc is usually then the suffix 2014-12-12 07:46:57 kernel seems to add .sign 2014-12-12 07:47:21 i just did: gpg --detach-sign $file 2014-12-12 07:47:28 which chreated .sig 2014-12-12 07:48:02 see e.g: ftp://ftp.funet.fi/pub/Linux/kernel/v3.x/linux-3.18.tar.sign 2014-12-12 07:48:12 hum 2014-12-12 07:48:23 kernel seems to sign the .tar; not the .xz and .gz 2014-12-12 07:48:55 fabled i added it: http://sprunge.us/HeOh but its stil there F2FS-fs (loop0): Can't find valid F2FS filesystem in 1th superblock 2014-12-12 07:50:26 hmm. sounds like that's then a config difference 2014-12-12 07:51:04 if the loop back image's fstype (squashfs) is not specified on mount time (like we don't as it used to be different format earlier) 2014-12-12 07:51:09 kernel will try to guess it 2014-12-12 07:51:19 and it can produce those errors which are safe to ignore 2014-12-12 07:51:31 vkrishn, ok. ncopa, is it fine with you? 2014-12-12 07:51:45 $ sprunge < x86_64/alpine-xen-3.1.0-x86_64.iso.asc 2014-12-12 07:51:45 http://sprunge.us/ZTTF 2014-12-12 07:52:15 looks better :) 2014-12-12 07:53:10 i suppose it does not make sense to publich both .sig and .asc 2014-12-12 07:53:13 vkrishn, btw: is lua binding working ? 2014-12-12 07:54:22 fabled i am sure we can ignore them, but like i wrote its clean image and i report it. 2014-12-12 07:55:51 fabled maybe to disable /dev/cdrom http://sprunge.us/VNRc 2014-12-12 07:56:28 forget it, source for all image are the same :D 2014-12-12 07:56:30 fcolista, I tested basic omega, [php|lua]-bindings , but redo when pushed to main 2014-12-12 07:57:35 do you think libxapian-dev should be pkg'ed ? 2014-12-12 07:57:56 vkrishn: do you intentionally edit the downloads pages css? http://sprunge.us/EJGO 2014-12-12 08:01:36 vkrishn: please rebase the css or send me a diff 2014-12-12 08:01:50 ok 2014-12-12 08:03:00 vkrishn, libxapian is a subpkg of xapian-core, and (i suppose) that xapian-core-dev contains -dev files for libxapian. Do you want also a subpkg for libxapian? 2014-12-12 08:03:47 I don't know, maybe ncopa could suggest, but seen libxapian-dev pkgs somewhere 2014-12-12 08:03:56 let me check if it was debian 2014-12-12 08:04:46 otherwise xapian-core-dev is ok 2014-12-12 08:06:08 barthalion: thanks for the website fix :) 2014-12-12 08:06:17 ncopa, i'm not able to fix build issue of v4l-utils. With this patch: http://sprunge.us/hPjV 2014-12-12 08:06:24 i've still this problem: http://pastebin.com/kh3W8dtk 2014-12-12 08:07:05 fabled: i now use 'asc' instead of 'sig': http://wwwtest.alpinelinux.org 2014-12-12 08:07:05 well, actually not "still". I've this other problem. 2014-12-12 08:07:12 please also check downloads page 2014-12-12 08:07:31 if those are ok i'll just merge it to master branch 2014-12-12 08:08:11 so xapian 2014-12-12 08:08:14 whats the status there? 2014-12-12 08:08:44 vkrishn, tested and reported it works fine. 2014-12-12 08:09:07 3 pkgs, xapian-{core,bindings,omega} 2014-12-12 08:09:11 was all 3 tested? 2014-12-12 08:09:48 not all bindings, not sure if I can 2014-12-12 08:09:55 ok 2014-12-12 08:10:03 but you did test some of them? 2014-12-12 08:10:16 lua, php 2014-12-12 08:10:24 but think I would test python when pushed to main 2014-12-12 08:10:29 ncopa in case you missed it: https://alpinelinux.org/ the middle "Download Latest Section" seems broken with some extra characters. 2014-12-12 08:10:50 :D 2014-12-12 08:11:09 cool :) 2014-12-12 08:11:20 can favicon be replaced here, http://git.alpinelinux.org/cgit/aports/log/ 2014-12-12 08:11:27 ncopa, please make https same as http. some browsers take directly to https if it's enabled 2014-12-12 08:12:21 fcolista: xapian-core looks ok 2014-12-12 08:13:20 bindings 2014-12-12 08:13:21 hum 2014-12-12 08:13:32 no support for lua5.2? 2014-12-12 08:13:41 i was kinda hoping that we could kill lua5.1 2014-12-12 08:13:52 and only support lua 5.2 and 5.3 2014-12-12 08:13:56 It has. I was wondering to enable it, but i had some troubles. 2014-12-12 08:14:02 also 2014-12-12 08:14:08 may be lua compat packaged 2014-12-12 08:14:08 site_perl? 2014-12-12 08:14:51 is this worng? PERL_LIB=/usr/lib/perl5/site_perl 2014-12-12 08:15:24 to my understanding yes 2014-12-12 08:16:22 since we are the packagers, it should go to 'vendorlib' 2014-12-12 08:16:42 sitelib is where stuff people compile themselves ends up 2014-12-12 08:16:47 site specific 2014-12-12 08:16:59 i always found confusing perl options regards lib... 2014-12-12 08:17:11 http://www.thecodingforums.com/threads/perl-vs-site-vs-vendor.898105/ 2014-12-12 08:17:38 -Dsiteprefix=/usr/local \ 2014-12-12 08:17:38 -Dsitelib=/usr/local/share/perl5/site_perl \ 2014-12-12 08:17:38 -Dsitearch=/usr/local/lib/perl5/site_perl \ 2014-12-12 08:17:52 yup, sitelib is under /usr/local 2014-12-12 08:18:34 which is where you put stuff you dont want the package manager to manage 2014-12-12 08:19:07 ok so are custom config or compilation 2014-12-12 08:19:16 yup 2014-12-12 08:19:18 this is what site is used for 2014-12-12 08:19:27 while vendor is the standar place for libs 2014-12-12 08:19:30 *standard 2014-12-12 08:19:58 for things managed by package manager yes 2014-12-12 08:20:11 in our case: /usr/share/perl5/vendor_perl 2014-12-12 08:20:20 cat APKBUILD | grep vend 2014-12-12 08:20:20 _vendorlib=/usr/share/perl5/vendor_perl 2014-12-12 08:20:21 _vendorarch=/usr/lib/perl5/vendor_perl 2014-12-12 08:20:21 -Dvendorprefix=/usr \ 2014-12-12 08:20:21 -Dvendorlib=$_vendorlib \ 2014-12-12 08:20:21 -Dvendorarch=$_vendorarch \ 2014-12-12 08:21:03 eval local perl -V:vendorlib 2014-12-12 08:21:14 $vendorlib 2014-12-12 08:21:28 thats how git does it 2014-12-12 08:23:30 is this variable enabled while compiling perl? 2014-12-12 08:23:38 Or when it is installed? 2014-12-12 08:23:39 yes 2014-12-12 08:23:47 $ perl -V:vendorlib 2014-12-12 08:23:47 vendorlib='/usr/share/perl5/vendor_perl'; 2014-12-12 08:24:23 this is what i've set on PERL_LIB 2014-12-12 08:24:31 rather than sitelib 2014-12-12 08:24:38 site_perl sorry 2014-12-12 08:24:41 they dont have a configure option for it? 2014-12-12 08:25:08 they have 2014-12-12 08:25:09 PERL perl interpreter 2014-12-12 08:25:09 PERL_INC Directory to include for XSUB,h and other Perl API headers 2014-12-12 08:25:09 PERL_ARCH Directory to install architecture-dependent perl files into 2014-12-12 08:25:09 PERL_LIB Directory to install architecture-independent perl files into 2014-12-12 08:25:10 PERL_SO for perl module extension 2014-12-12 08:25:50 i mean: ./configure --with-perl=vendor ... or similar 2014-12-12 08:26:14 ah 2014-12-12 08:26:19 no, doesn't seems so 2014-12-12 08:26:20 --with-perl enable Perl bindings 2014-12-12 08:26:36 maybe is undocumented? 2014-12-12 08:26:46 gotta check the configure file 2014-12-12 08:34:29 ncopa, can this be applied, just validator fixes https://dpaste.de/So6A/raw 2014-12-12 08:36:21 i added it to wwwtest branch already 2014-12-12 08:36:31 i first push to wwwtest 2014-12-12 08:36:55 when once that is tested and verified to be ok i merge it to master 2014-12-12 08:37:04 hhum 2014-12-12 08:37:08 maybe i should do the opposite 2014-12-12 08:37:17 create a 'production' branch 2014-12-12 08:37:42 people tend to use git master for development 2014-12-12 08:37:58 ok 2014-12-12 08:38:07 let me fix the site git repo branches 2014-12-12 08:38:19 ncopa, i wonder if we should revert the change to 'ld.gold' as default 2014-12-12 08:38:33 morning 2014-12-12 08:38:47 fabled: broken and not fixable? 2014-12-12 08:39:02 https://sourceware.org/bugzilla/show_bug.cgi?id=17699 2014-12-12 08:39:08 broke libstdc++ on x86 2014-12-12 08:39:18 so, patch send againt wwwtest branch ? 2014-12-12 08:39:52 vkrishn: please wait a bit 2014-12-12 08:40:00 i'll make git master the development branch 2014-12-12 08:40:05 or dev branch, quite common 2014-12-12 08:40:07 and have a 'production' branch 2014-12-12 08:40:43 fabled: yeah, maybe revert gold as default til that issue is resolved 2014-12-12 08:41:11 gitworkflows has some nice methods 2014-12-12 08:41:18 both ld.bfd and ld.gold have a huge number of bugs open 2014-12-12 08:41:38 not sure how many 'critical' bugs in gold 2014-12-12 08:41:55 bfd linker does support gazillion more target architeures, a.out and other stuff too 2014-12-12 08:42:03 gold is really ELF-only linker 2014-12-12 08:45:22 hmm 2014-12-12 08:47:25 i'll revert to ld.bfd 2014-12-12 08:47:32 seems there's few other nasty gold bugs too 2014-12-12 08:47:58 should i keep gold enabled still? 2014-12-12 08:48:52 ncopa, ^^^ ? 2014-12-12 08:49:22 there is no aporttest branch ;) 2014-12-12 08:50:32 or aportstest , eehh 2014-12-12 08:50:46 fabled: i'm ok to keep gold available for those who wants to test 2014-12-12 08:50:56 some apps might detect and use it by default 2014-12-12 08:51:04 e.g. qt seems to do it 2014-12-12 08:51:10 hum 2014-12-12 08:51:18 can we ship it in subpackage? 2014-12-12 08:51:26 binutils-gold? 2014-12-12 08:51:28 would be possible 2014-12-12 08:51:33 i think some do that 2014-12-12 08:51:43 though i'm not sure how to do the split 2014-12-12 08:52:06 well 2014-12-12 08:52:11 vkrishn, barthalion: you can use git master for wwwtest now 2014-12-12 08:52:11 probably just moving ld.gold to it 2014-12-12 08:52:18 yeah 2014-12-12 08:52:37 ok 2014-12-12 08:54:30 http://www.alpinelinux.org/forum is broken 2014-12-12 08:55:00 ok 2014-12-12 08:55:05 when you get to it from here http://www.alpinelinux.org/packages/ 2014-12-12 08:55:08 i think its time to move to forums.alpinelinux.org 2014-12-12 08:55:20 as is http://www.alpinelinux.org/login 2014-12-12 08:55:43 also as fabled pointed out, its good if https:// and http:// point to same location 2014-12-12 08:55:59 so i think we will have to move www.a.o to forum.a.o 2014-12-12 08:56:01 or so 2014-12-12 09:00:28 can you remind me the url to clone for the site? 2014-12-12 09:00:51 http://git.alpinelinux.org/cgit/alpine-mksite ? 2014-12-12 09:01:20 there is a list on the bottom of each repository 2014-12-12 09:01:26 yeah 2014-12-12 09:01:31 http://git.alpinelinux.org/cgit/alpine-mksite at the bottom 2014-12-12 09:01:38 says: Clone 2014-12-12 09:01:38 git://git.alpinelinux.org/alpine-mksite 2014-12-12 09:01:50 but unless you have push access, use http:// or git:// 2014-12-12 09:02:08 and the link you posted is actually working git remote 2014-12-12 09:02:29 ncopa, http://sprunge.us/YcBL is that ok looking split? 2014-12-12 09:03:12 looks good to me 2014-12-12 09:03:35 barthalion: ok, and the repos are here http://git.alpinelinux.org/ . got it. 2014-12-12 09:04:34 thanks for that. i'm very new to all this and it hasn't become second nature yet 2014-12-12 09:05:10 ncopa, https://dpaste.de/fYLk/raw , improved mobile devices layout 2014-12-12 09:07:38 ARGH 2014-12-12 09:08:53 sorry 2014-12-12 09:08:56 i neeed coffee 2014-12-12 09:09:24 any thoughts on renaming alpine-mksite to "fractal" ;-) 2014-12-12 09:10:08 or "snowflakes" 2014-12-12 09:12:55 and apk to tauromachia 2014-12-12 09:12:57 ACTION runs 2014-12-12 09:23:18 heh, i used to type apt-get install in alpinelinux, but now i'm typing apk get in debian :) 2014-12-12 09:33:55 ncopa, should i also push http://sprunge.us/SGVZ ? 2014-12-12 09:36:19 yeah 2014-12-12 09:38:11 vkrishn: 2014-12-12 09:38:13 table tr td:first-child { 2014-12-12 09:38:13 white-space: nowrap; 2014-12-12 09:38:13 } 2014-12-12 09:38:18 whats thats supposed to do? 2014-12-12 09:38:58 on <320 px Alpine News data column '2014-12-10' wraps 2014-12-12 09:39:17 why dont we do that everywhere? 2014-12-12 09:39:50 on wider screen is less likely to wrap 2014-12-12 09:39:51 but ok, either 2014-12-12 09:40:07 but remember this is global set, so may effect other tables on other pages 2014-12-12 09:40:35 best add a class to "Alpine News" and "Latest Development" 2014-12-12 09:41:14 or jut date? 2014-12-12 09:41:18 just* 2014-12-12 09:41:19 ok 2014-12-12 09:42:03 wait date =?? 2014-12-12 09:42:40 i mean