2013-08-01 06:09:07 morning 2013-08-01 06:09:32 seems like its trivial to implement NETDOWN=no like functionality 2013-08-01 06:10:00 but it is even easier with KEEP_IFACE_UP="eth0 eth3" 2013-08-01 06:39:28 files from v2.5.4-173-g3575990 uploaded 2013-08-01 06:39:40 files from v2.5.4-173-g3575990 uploaded 2013-08-01 08:39:54 hi 2013-08-01 08:52:07 ncopa: when will be the new mirror in 2.6? 2013-08-01 08:52:27 shafire: hi 2013-08-01 08:52:32 which mirror? 2013-08-01 08:52:52 esslingen 2013-08-01 08:53:07 ok 2013-08-01 08:53:20 i only added it to edge so it will be added to v2.7 2013-08-01 08:53:46 can you create an issue on bugs.a.o, with target 2.6.3 2013-08-01 08:53:52 and ask for adding it? 2013-08-01 08:53:54 so i dont forget 2013-08-01 08:53:57 ok 2013-08-01 08:54:04 i'd like to make new v2.6 release next week 2013-08-01 08:59:21 What's the 2.7 roadmap? 2013-08-01 08:59:32 because of the mirror json file 2013-08-01 09:08:46 I see: 2013-11-01 2013-08-01 09:38:12 we can add it to next v2.6, np 2013-08-01 09:41:22 :-) 2013-08-01 11:52:06 i will add a config option for abuild 2013-08-01 11:52:17 CLEANUP="srcdir pkgdir deps" 2013-08-01 11:52:34 after successful build it will remove srcdir pkgdir and uninstall deps 2013-08-01 11:52:50 that way you can specify what you want abuild to clean up after successful build 2013-08-01 11:53:53 i wonder if we should have something for failures too 2013-08-01 11:54:17 like, ERROR_CLEANUP="deps" 2013-08-01 11:54:25 that will uninstall deps on build failure 2013-08-01 11:54:41 on a dev box it might make sense to not uninstall the makedepends on failure 2013-08-01 11:54:55 but on build servers we want uninstall them 2013-08-01 12:26:20 fabled: of for a release of abuild? 2013-08-01 12:26:49 fabled: is this ok? http://sprunge.us/DZdI 2013-08-01 12:27:01 maybe should add arm/musl example there too 2013-08-01 12:29:42 yes looks ok 2013-08-01 12:29:43 oh 2013-08-01 12:29:47 ncopa, update newapkbuild 2013-08-01 12:29:55 with the ./configure changes 2013-08-01 12:30:12 true 2013-08-01 13:10:31 fabled: I think I have found a bug (or feature?) in uclibc's ld.so 2013-08-01 13:10:43 LD_LIBRARY_PATH=/path/to/dir 2013-08-01 13:10:44 works 2013-08-01 13:10:55 but LD_LIBRARY_PATH=:/path/to/dir does not 2013-08-01 13:10:59 note the : 2013-08-01 13:11:13 it breaks a waf script 2013-08-01 13:35:31 all bugs are features in uclibc tbh 2013-08-01 13:36:04 and I don't think that LD_LIBRARY_PATH should start with : 2013-08-01 13:39:22 barthalion: correct. i fixed the configure script 2013-08-01 13:39:42 how will you decide to make v3? 2013-08-01 13:40:09 shafire: kernel fixes 2013-08-01 13:40:14 and not too often 2013-08-01 13:40:25 i am trying to make approx a stable release every month 2013-08-01 13:40:35 I guess that glibc is just more failproof for stuff like that, I wouldn't blame uclibc too much 2013-08-01 13:40:51 well, it breaks configure scripts 2013-08-01 13:41:10 trivial fix though: http://sprunge.us/aFJR 2013-08-01 13:47:36 shafire: v3.0 will come after v2.7 2013-08-01 13:47:44 kaniini: why? 2013-08-01 13:48:07 shafire: we are using ARM as kind of a trial run on some major changes 2013-08-01 13:48:21 will you support arm64 in future? :-) 2013-08-01 13:48:25 shafire: probably. 2013-08-01 13:48:37 shafire: but, basically, we're ditching uclibc after 2.7 2013-08-01 13:48:49 by? 2013-08-01 13:48:53 shafire: musl-libc 2013-08-01 13:49:38 shafire: we decided, basically, it's better for alpine if our libc is actively maintained instead of us having to patch it constantly :P 2013-08-01 13:49:55 :-) 2013-08-01 13:49:56 http://www.etalabs.net/compare_libcs.html 2013-08-01 13:53:20 i doubt we will be able to ditch uclibc after v2.7 2013-08-01 13:53:22 would be nice 2013-08-01 13:53:25 but we'll see 2013-08-01 13:53:36 yeah, we might have to do a 2.8 2013-08-01 13:53:52 but there will probably be a 3.0 branch along side 2.8 if there is one 2013-08-01 13:54:10 since the next step will obviously be x86-musl/x86_64-musl builders 2013-08-01 13:54:39 but basically, my point is 2013-08-01 13:54:50 3.0 will be the switch to musl across all archs 2013-08-01 13:55:01 as 2.0 was the switch to nptl 2013-08-01 13:57:21 correct 2013-08-01 13:58:00 fabled: i think we have all features for a new abuild release 2013-08-01 14:01:01 I will help you as good as I can :-) 2013-08-01 14:08:10 ncopa, great! 2013-08-01 14:08:47 there are actuanlly some nice enhancements 2013-08-01 14:09:01 instead of abuild installdeps you now can do: abuild deps 2013-08-01 14:09:03 its shorter 2013-08-01 14:09:12 and abuild rootpkg actually works 2013-08-01 14:41:22 hmpf 2013-08-01 14:41:31 i think there are bugs in samba build scripts or waf 2013-08-01 14:41:44 seems like building ends up in a deadlock once in a while 2013-08-01 22:21:41 oops 2013-08-01 22:21:46 just pretend that never happened 2013-08-01 22:22:20 files from v2.5.4-174-gf8dedb2 uploaded 2013-08-01 22:23:34 files from v2.5.4-175-g44794d2 uploaded 2013-08-01 22:23:43 files from v2.5.4-175-g44794d2 uploaded 2013-08-01 22:23:51 files from v2.4.11-72-g52c9187 uploaded 2013-08-01 22:23:57 files from v2.4.11-72-g52c9187 uploaded 2013-08-01 22:24:16 files from v2.3.6-255-g876e7ab uploaded 2013-08-01 22:24:23 files from v2.3.6-255-g876e7ab uploaded 2013-08-01 22:26:26 files from v2.2.3-234-g7afcaa0 uploaded 2013-08-01 22:26:36 files from v2.2.3-234-g7afcaa0 uploaded 2013-08-02 09:43:39 phew 2013-08-02 09:43:40 ok 2013-08-02 09:44:02 i have looked over the 3.10.4 and 3.4.55 releases 2013-08-02 09:44:13 am picked almost all of the relvant commits 2013-08-02 09:44:18 for our v2.6 kernel 2013-08-02 09:44:23 the 3.9 is EOL 2013-08-02 09:44:46 I am testbuilding it now and if it boots I'll push it to github 2013-08-02 09:44:56 where i nowdays trac our stable grsec kernels 2013-08-02 09:45:16 this is basicalyy 3.9.12 2013-08-02 09:45:27 i wonder if I should tag it as 3.9.12 2013-08-02 09:45:47 because that is basically what it is 2013-08-02 09:47:20 we should post info that we're taking over 3.9 maintenance then 2013-08-02 09:47:25 on lkml 2013-08-02 09:47:26 well 2013-08-02 09:47:42 i dunno for how long time i can do that 2013-08-02 09:48:00 IMO 3.9.11.1 sounds better, or just bump pkgrel 2013-08-02 09:48:17 i think i'll just bump pkgrel 2013-08-02 09:48:31 oh and actually 2013-08-02 09:48:40 its not a vanilla 3.9.12 2013-08-02 09:48:44 its 3.9.12-grsec 2013-08-02 09:48:56 should be fairly trivial though, to make it a 3.9.12 vanilla 2013-08-02 10:28:07 hmm 2013-08-02 10:28:17 oh, ragel is in testing still 2013-08-02 10:58:56 royger: hi 2013-08-02 10:59:20 i am looking over the commits in linux-3.10.y and linux-3.4.y 2013-08-02 10:59:36 ncopa: hello 2013-08-02 10:59:49 do you know if we want this commit for our v2.6 (kerle 3.9.y): 2013-08-02 10:59:56 https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?h=linux-3.10.y&id=2d06fa0f825ec1d7943a3bcceecbd85aa82ef27d 2013-08-02 11:00:38 "A recent change in the resume code uses the softirq based delayed clock_was_set to support Xens resume mechanism." 2013-08-02 11:00:45 but it does not say what commit 2013-08-02 11:00:54 so i dont know if 3.9 kernel is affected 2013-08-02 11:02:07 acutally 2013-08-02 11:02:10 i know 2013-08-02 11:02:24 i just figured out that they added it to 3.4.y 2013-08-02 11:02:27 with 3.4.55 2013-08-02 11:02:30 so we want it 2013-08-02 11:04:04 I'm not sure about which change they are referring to, but I guess it doesn't hurt to have the commit anyway 2013-08-02 11:04:30 normally it does not hurt, yes 2013-08-02 11:04:33 sometimes it does 2013-08-02 11:04:50 it seems like they just move the call to clock_was_set to a kworker in order to prevent calling it from softirq context 2013-08-02 11:04:52 so i check the 'when introduced commit' 2013-08-02 11:05:08 but thye pushed it to 3.4.y, so we want it 2013-08-02 11:05:49 ok 2013-08-02 11:05:50 super 2013-08-02 11:06:22 then my list of questionable commits got shorter :) 2013-08-02 11:06:48 I'm not much familiar with linux timer code, but I would say it looks quite harmless 2013-08-02 11:07:00 i added it 2013-08-02 11:07:02 thanks! 2013-08-02 11:07:10 thanks for looking at it 2013-08-02 11:07:21 np :) 2013-08-02 11:07:36 sorry for not being available during the week, had problems with my internet connection 2013-08-02 11:07:48 i will try make a v2.6.3 release next week 2013-08-02 11:07:51 np 2013-08-02 11:09:02 fabled: here are some notes of commits i'd like you to have a look at: https://dpaste.de/5GIek/ 2013-08-02 11:09:38 i dont remember if 'gre: Fix MTU sizing check for gretap tunnels' fails to apply 2013-08-02 11:09:47 but i noted it as 'ask fabled to look at it' 2013-08-02 11:09:53 ext4 did not apply 2013-08-02 11:10:43 virto_net applies but fails on build. do we want that fix? we backport it properly? 2013-08-02 11:10:55 virtio_net was not added to 3.4.y 2013-08-02 11:11:36 then there are 7 commits that were added to 3.4.y that was nott added to 3.10.y 2013-08-02 11:12:26 and afaik the 'upstream commit' is not 'git merge-base --is-ancestor' in 3.10 2013-08-02 11:14:41 the exact list of commits in 3.4.y (3.4.54->3.4.55) that does not appear in 3.10.3->3.10.4, but has "upstream commit" in 3.10.y: https://dpaste.de/uSsJp/ 2013-08-02 11:14:42 ncopa, https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?h=linux-3.10.y&id=6afbcb59964ec9dc3dd63ed96a9e2cda31f8a072 will fail to apply as it's regression by ip_tunnels rewrite for 3.10.x 2013-08-02 11:14:50 so that should be dropped 2013-08-02 11:15:17 good, thanks! 2013-08-02 11:16:09 not sure about the others with quick look 2013-08-02 11:16:18 like me then :) 2013-08-02 11:16:23 "not sure" :) 2013-08-02 11:16:26 thats why i ask 2013-08-02 11:24:31 ok, i suppose we just let them out til someone asks for them 2013-08-02 12:34:42 fabled: arm buildbox running? 2013-08-02 12:53:37 clandmeter: i think it's the "algitbot" user 2013-08-02 12:54:20 jbilyk_: what do you mean? 2013-08-02 12:54:22 and hi 2013-08-02 12:54:27 clandmeter: hi 2013-08-02 12:54:50 i don't think that the buildozer on the arm builder is running in this channel as "build-*", it's the algitbot nick 2013-08-02 12:55:00 hasn't been changed to buildarm-edge or similar yet 2013-08-02 12:56:06 yes, but i guess it first need to build aports tree before its going to active. 2013-08-02 12:56:20 and according to what ive read, thats going to take a looong time... 2013-08-02 12:58:18 clandmeter: yeah...at looks like it 2013-08-02 12:58:22 at the moment* 2013-08-02 13:22:16 ncopa, ^^^seems some isseu in py-tevent ? 2013-08-02 13:23:00 clandmeter, arm vserver is up, but not doing much now. it still has some issues. 2013-08-02 13:23:11 mostly qemu-user cannot reliably run multithreaded programs 2013-08-02 13:23:58 fabled: is that something you can/want to fix? 2013-08-02 13:24:29 clandmeter, it sounds like non-trivial to fix. i'm thinking rather to get some ODROID-X2 units. 2013-08-02 13:24:39 or use qemu-system (which is a lot slower, but would work) 2013-08-02 13:25:56 hmm, I have a concern 2013-08-02 13:26:05 today i've been doing other projects. should have some time for arm again next week. 2013-08-02 13:26:08 shouldn't -dev packages depend on the exact versions of their parents? 2013-08-02 13:26:17 hl, new abuild does that. 2013-08-02 13:26:20 fabled: You want to purchase them yourself? 2013-08-02 13:26:25 fabled: where do I get 'new abuild'? 2013-08-02 13:26:33 hl: yes, its in git 2013-08-02 13:26:40 ncopa: what version ETA? 2013-08-02 13:26:55 hl, http://git.alpinelinux.org/cgit/abuild/ 2013-08-02 13:26:59 2.7? 2.8? 2013-08-02 13:26:59 it was supposed to be released today 2013-08-02 13:27:04 hl: 2.7 2013-08-02 13:27:08 ah, great. 2013-08-02 13:27:08 and edge ofc 2013-08-02 13:27:11 http://git.alpinelinux.org/cgit/abuild/commit/?id=ad50945d80451fccbbdc474ae84234b893834024 2013-08-02 13:27:22 ncopa, i have one more wish for abuild 2013-08-02 13:27:24 I just did apk add dovecot@edge to get 2.2.4, and was alarmed to see -dev was still @main, 2.2.1. 2013-08-02 13:27:41 I'll just pin the rest for now. 2013-08-02 13:28:07 yes, you'll have to pin til alpine v2.7 2013-08-02 13:28:17 i'll push new abuild today if i reach 2013-08-02 13:28:26 then all packages needs to be rebuilt... 2013-08-02 13:28:38 we do rebuild all packages when we make new stable releases 2013-08-02 13:28:54 so edge packages and v2.7 will differ in that regard 2013-08-02 13:31:03 ncopa, oh, it's apk-tools issue. seems --quiet --simulate --interactive will still result in question asked 2013-08-02 13:31:08 will fix apk-tools 2013-08-02 13:32:02 i dont use --interactive 2013-08-02 13:32:20 i use --simulate if in doubt 2013-08-02 13:32:43 fabled: did you see the new odroid product? seems announced now. 2013-08-02 13:32:51 clandmeter, oh? 2013-08-02 13:32:56 octo 2013-08-02 13:33:11 ODROID-XU 2013-08-02 13:33:33 usb3 2013-08-02 13:34:56 stil only 10/100 network. 2013-08-02 13:35:51 fabled: was tehre any abuild feature you wanted? 2013-08-02 13:35:54 or I'll tag it now 2013-08-02 13:37:17 ncopa, no it's apk issue. i have --interactive on by default (/etc/apk/interactive) 2013-08-02 13:38:07 clandmeter, cool. odroid-xu looks very nice 2013-08-02 13:38:25 did you see what xu stand for? lol 2013-08-02 13:38:36 XU means eXtreme + Ultimate ! 2013-08-02 13:38:47 i wonder what the next version will be.. 2013-08-02 13:39:00 lol 2013-08-02 13:39:07 XUNG 2013-08-02 13:39:16 extreme ultimate next gen 2013-08-02 13:39:55 don't forget super, hiper, etc 2013-08-02 13:40:09 fabled: i dont mind contributing to buy one. maybe some others are interested as well. 2013-08-02 13:41:06 might be good. do you know when odroid-xu ships? 2013-08-02 13:41:19 sept 10 2013-08-02 13:41:44 The first shipment will be started from 10th of September. If you purchase ODROID-XU / XU+E and/or its accessories together, the shipment will be in September. 2013-08-02 13:42:56 the base package should be enough. with usb3, no need for eMMC. 2013-08-02 13:43:25 179 dollar inc shipping ex tax. 2013-08-02 13:43:30 Samsung’s Exynos 5410 chip is an 8-core processor with 4 ARM Cortex-A15 cores paired with 4 ARM Cortex-A7 cores and PowerVR SGX544MP3 graphics. At any given time, only one set of CPU cores or the other is actually active. 2013-08-02 13:43:33 8cores! 2013-08-02 13:43:55 yes, but the 5410 has some issues afaik 2013-08-02 13:44:11 its prefered to get the 5420 2013-08-02 13:44:27 but we dont have much choise.. 2013-08-02 13:45:37 i wonder if there are any similar, but without gpu 2013-08-02 13:45:42 for compile server 2013-08-02 13:45:54 heh 2013-08-02 13:46:12 maybe we should have set up a NAS and iscsi 2013-08-02 13:46:32 so those builders could store data 2013-08-02 13:47:25 you will need a faster nic 2013-08-02 13:47:30 oh true 2013-08-02 13:47:32 its only 100 2013-08-02 13:47:37 yup 2013-08-02 13:47:48 but you could use usb 1G nic 2013-08-02 13:47:49 and in previous versions it was via USB host interface 2013-08-02 13:47:56 but at least it has USB3.0 2013-08-02 13:48:14 with a simple ssd, this should work 2013-08-02 13:55:35 hello guys, what version of Uclibc does alpine-linux use ? 2013-08-02 13:56:13 rendar_: 0.9.33.2 with our own patchset 2013-08-02 13:56:15 rendar_: http://git.alpinelinux.org/cgit/uClibc-alpine 2013-08-02 13:56:39 most of the patches are in upstream git 2013-08-02 13:57:09 good! 2013-08-02 14:00:59 what version of gcc included ? 2013-08-02 14:02:03 4.8.1 in edge, probably 4.7.x in latest stable 2013-08-02 14:02:35 cool 2013-08-02 14:03:18 going to try it for uClibc development environment 2013-08-02 14:03:22 thanks! 2013-08-02 14:03:45 np :) 2013-08-02 14:09:51 ok 2013-08-02 14:09:56 here comes abuild-2.16 2013-08-02 19:37:27 hello guys, I was very surprised, when I "apk add gcc" and stdio.h and all .h devel files were not there, shouldnt uclibc-dev be a default dependency for gcc ? 2013-08-02 19:40:09 is there a gcc-dev rendar_ ? 2013-08-02 19:40:27 no 2013-08-02 20:01:36 rendar_: i suggest adding alpine-sdk 2013-08-02 20:01:49 it has all of the necessary packages included as part of that virtual package 2013-08-02 20:18:17 got "Can't modify application's text section; use the GCC option -fPIE for position-independent executables." is that familiar ? 2013-08-02 20:18:55 well, why won't you use -fPIE? 2013-08-02 20:19:37 because it is used inside of perl CPAN ... and don't even know where 2013-08-02 20:20:46 I'm sure we have a tool to package perl modules, or at least automatize it 2013-08-02 20:21:37 apkgbuild-cpan or something like that 2013-08-02 20:22:00 apkbuild-cpan 2013-08-02 20:22:45 I'm doing too much Arch packaging nowadays, I mix stuff 2013-08-02 20:23:42 maybe, but it is a part software that is not related to local perl 2013-08-02 20:24:08 i want to compile "App-Staticperl", and is a reason why I installed Alpine linux at all 2013-08-02 20:24:32 oh god, CVS 2013-08-02 20:24:56 I can't believe that someone is using it yet 2013-08-02 20:25:22 it is not somebody, it is Marc Lehmann ! 2013-08-02 20:26:48 which script is supposed to build an executable? 2013-08-02 20:27:53 ./bin/staticperl install, wait for ~10 minutes... error while installing perl modules, in CPAN ovbiously 2013-08-02 20:28:59 have you tried adding -fPIE to PERL_CCFLAGS? 2013-08-02 20:30:24 not yet, but I have found somebody else having same problem particularly in Alpine linux 2013-08-02 20:35:59 is it possible to rebuild uclibc inside of alpine linux, I think it is related ro uClibc compile flag 2013-08-02 20:36:02 ? 2013-08-02 20:36:47 let me reproduce the issue first 2013-08-02 20:39:52 lets just try this flag for gcc, i think there should be file for default gcc flags somewhere, don't you know where ? 2013-08-02 20:40:16 you pointed me to this file :) 2013-08-02 20:40:47 ah, i will just make wrapper for gcc 2013-08-02 20:41:56 there is PERL_CCFLAGS variable in bin/staticperl 2013-08-02 20:43:08 it does not help 2013-08-02 20:43:24 weird 2013-08-02 20:43:54 probably because it does nothing, because CPAN uses it's own flags inside 2013-08-02 20:49:43 with gcc wraper and -fPIE enabled i get same error 2013-08-02 20:50:10 I need to get some sleep, I'll take a look tommorow 2013-08-02 20:50:19 okay, thanks 2013-08-02 21:01:31 barthalion: problem solved, reason was in manually compilled gzip (from source), recompilling it with -fPIE helped 2013-08-02 21:02:55 and i did it because Alpine's gzip barks on "-q" flag 2013-08-02 21:03:51 well, you could just install gzip 2013-08-02 21:03:57 you probably used busybox one 2013-08-02 21:04:35 apk add gzip does nothing, so I use Alpine's one 2013-08-02 21:05:09 -q works fine here 2013-08-02 21:05:17 hmm 2013-08-02 21:05:53 while busybox gzip -q foobar indeed fails 2013-08-02 21:07:31 so, where can I get gzip ? 2013-08-02 21:07:43 apk add gzip should be enough 2013-08-02 21:07:49 ls -lh /bin/gzip 2013-08-02 21:08:09 nope, look: 2013-08-02 21:08:31 apk add gzip 2013-08-02 21:08:32 (1/1) Installing gzip (1.5-r0) 2013-08-02 21:08:32 Executing busybox-1.21.1-r0.trigger 2013-08-02 21:08:57 so you just installed it 2013-08-02 21:09:32 really, strange 2013-08-02 21:10:29 anyway, good to hear that everything works now :) 2013-08-02 21:10:30 I think it is time to sleep for me too, I start making mistakes :) 2013-08-02 21:10:31 good night 2013-08-02 21:10:49 good night 2013-08-04 22:37:41 rnalrd: do you mind if we stick to 3.10 (new lts) with linux-virt-grsec? I'll try to backport relevant grsec updates 2013-08-05 05:08:30 ncopa, "Also worth noting: Greg Kroah-Hartman has announced that 3.10 will be the next long-term supported kernel." 2013-08-05 06:06:23 morning 2013-08-05 06:06:24 oh! 2013-08-05 06:06:36 3.10 2013-08-05 06:06:46 then i think we'll stick with 3.10 for v2.7 too 2013-08-05 06:06:57 either that or we upgrade v2.6 to 3.10 2013-08-05 06:07:23 barthalion, i'm ok with that 2013-08-05 06:07:43 in fact, i think i'd like both v2.6 and v2.7 stay on 3.10 2013-08-05 06:07:49 since looks like ncopa is going to push 3.10 for linux-grsec i would just wait for him :) 2013-08-05 06:07:50 will save us for lots of pain 2013-08-05 06:07:55 i will 2013-08-05 06:07:57 no doubt 2013-08-05 06:08:04 and import the patches 2013-08-05 06:08:49 fabled: do you think that it is realistic to upgrade the v2.6 to kernel 3.10? 2013-08-05 06:30:34 ncopa, need to test first. but yes, i think we can do it after some testing @edge 2013-08-05 06:30:52 super 2013-08-05 06:31:21 the question is, we do it for v2.6.3? or we calm down and aim for v2.6.4? 2013-08-05 06:31:30 hm 2013-08-05 06:31:35 with the virt kernel 2013-08-05 06:31:36 depends when you want a release 2013-08-05 06:31:42 i want release tomorrow 2013-08-05 06:31:56 well - i think i can upgrade dev-vpnc today 2013-08-05 06:32:27 i could test my dmvpn stuff then 2013-08-05 06:32:53 i wonder if we for v2.7 should remove the paravrit stuff from kernel 2013-08-05 06:32:59 but it might be still better to wait a bit more 2013-08-05 06:33:10 ok 2013-08-05 06:33:15 i'll build it now 2013-08-05 06:33:18 you test it 2013-08-05 06:33:24 yes, i'd be careful to change config in -stable 2013-08-05 06:33:43 the config will not change for stable 2013-08-05 06:34:36 where can i read about 3.10 beeing lts? 2013-08-05 06:35:31 http://www.kroah.com/log/blog/2013/08/04/longterm-kernel-3-dot-10/ 2013-08-05 06:37:10 NETLINK: mmaped IO (NETLINK_MMAP) [N/y/?] (NEW) ? 2013-08-05 06:37:10 CONFIG_NETLINK_MMAP: 2013-08-05 06:37:10 This option enables support for memory mapped netlink IO. This 2013-08-05 06:37:10 userspace. 2013-08-05 06:37:10 reduces overhead by avoiding copying data between kernel- and 2013-08-05 06:37:11 If unsure, say N. 2013-08-05 06:37:14 sounds useful 2013-08-05 06:37:28 i wonder if i should disable new stuff for now 2013-08-05 06:37:33 yes, i might use it eventually in opennhrp 2013-08-05 06:37:33 for -stable 2013-08-05 06:37:42 ok 2013-08-05 06:37:47 enable it then 2013-08-05 06:37:54 though - it requires support in userspace. not sure how much stuff support it. 2013-08-05 06:38:07 very little i suppose 2013-08-05 06:38:24 but i wonder if i should disable it for -stable and then later re-enable it for edge 2013-08-05 06:38:27 for v2.7 2013-08-05 06:38:59 i'd go ahead and just enable it. 2013-08-05 06:39:32 even though it's new api; it's fundamentally just peformance improvement 2013-08-05 07:38:40 fabled: do we want CONFIG_CRYPTO_CMAC enabled for the v2.6 stable? 2013-08-05 07:38:42 CONFIG_CRYPTO_CMAC: 2013-08-05 07:38:43 Cipher-based Message Authentication Code (CMAC) specified by 2013-08-05 07:38:43 The National Institute of Standards and Technology (NIST). 2013-08-05 07:38:43 https://tools.ietf.org/html/rfc4493 2013-08-05 07:38:43 http://csrc.nist.gov/publications/nistpubs/800-38B/SP_800-38B.pdf 2013-08-05 07:39:16 seems like a new option 2013-08-05 07:44:00 umm, enable for edge. -stable if you fell like it. 2013-08-05 07:49:49 i have disabled practically everything new for -stable 2013-08-05 07:49:59 including cmac 2013-08-05 07:50:13 the only thing i enabled was 2013-08-05 07:50:14 * NETLINK: mmaped IO (NETLINK_MMAP) [N/y/?] (NEW) y 2013-08-05 07:50:14 * NETLINK: socket monitoring interface (NETLINK_DIAG) [N/m/y/?] (NEW) m 2013-08-05 07:59:28 ok 2013-08-05 07:59:36 sounds ok to me 2013-08-05 08:03:02 fabled, the 6 backport patches applied. do we still need them? http://sprunge.us/OeTU 2013-08-05 08:03:37 for 3.10 you mean? 2013-08-05 08:03:41 yes 2013-08-05 08:04:02 i need to look through logs. i think 3.11 should have them all, and many are in 3.10 too; so need to double check 2013-08-05 08:11:02 ncopa, i think those all are still needed with 3.10 2013-08-05 08:11:08 ok 2013-08-05 08:11:13 i'll keep them 2013-08-05 08:11:15 they seem to have went in for 3.11 2013-08-05 08:11:17 and we test 2013-08-05 09:26:31 localhost:~# uname -a 2013-08-05 09:26:32 Linux localhost 3.10.4-0-grsec #1-Alpine SMP Mon Aug 5 10:59:50 CEST 2013 x86_64 2013-08-05 09:26:32 Linux 2013-08-05 09:37:21 #1 2013-08-05 09:37:26 *g* 2013-08-05 09:37:31 funny bot-trigger ;-) 2013-08-05 09:39:19 yeah 2013-08-05 09:39:40 i wonder if i should filter out issues < 100 2013-08-05 09:39:55 or even 10 2013-08-05 09:39:57 #10 2013-08-05 09:41:16 #9 2013-08-05 09:42:07 ticket 1 2013-08-05 09:42:12 issue 1 2013-08-05 09:42:30 issue 1 2013-08-05 09:42:33 #1 2013-08-05 09:42:36 #9 2013-08-05 09:42:39 #10 2013-08-05 09:43:01 should reduce false positives 2013-08-05 11:13:12 2 errors; 2527 MiB in 995 packages 2013-08-05 11:13:26 fabled: how do i know what pkgs has errors? 2013-08-05 11:14:32 "apk fix --simulate" 2013-08-05 11:15:08 and apk fix alone fixed it 2013-08-05 11:15:11 yes 2013-08-05 11:15:28 ncopa-desktop:~/alpine-iso$ sudo apk fix --simulate 2013-08-05 11:15:28 (1/2) Reinstalling tdb-libs (1.2.12-r0) 2013-08-05 11:15:28 (2/2) Reinstalling tevent (0.9.18-r0) 2013-08-05 11:15:38 i think there was some file conflict 2013-08-05 11:15:59 there is no way now to know what that was about? 2013-08-05 11:16:04 i have an idea thought... 2013-08-05 11:26:46 fcolista: on #eng: http://pastebin.ita.wtbts.net/dc8ad9a3 2013-08-05 11:42:46 hum 2013-08-05 11:42:54 i have noticed that time is wrong on my laptop 2013-08-05 12:16:39 files from v2.3.6-256-g6e323da uploaded 2013-08-05 12:16:39 files from v2.3.6-256-g6e323da uploaded 2013-08-05 12:17:33 files from v2.5.4-176-gd8160ad uploaded 2013-08-05 12:17:40 files from v2.5.4-176-gd8160ad uploaded 2013-08-05 12:18:04 files from v2.4.11-73-ga78dcc7 uploaded 2013-08-05 12:18:12 files from v2.4.11-73-ga78dcc7 uploaded 2013-08-05 13:10:10 looks like they have fixes for 3.10 in spl/zfs git 2013-08-05 13:10:22 but i think I'll wait for the 0.6.2 release 2013-08-05 13:10:24 lazy 2013-08-05 13:52:18 fabled: did you test nhrp/ipsec with the 3.10 kernel? 2013-08-05 13:52:22 not yet 2013-08-05 13:52:32 i installed the kernel - but didn't reboot yet 2013-08-05 13:52:55 rebooting now 2013-08-05 13:52:58 ok 2013-08-05 13:53:04 i think we have an issue with chronyd 2013-08-05 13:53:14 it does not set the time for me anymore on bootup 2013-08-05 13:53:16 not sure why 2013-08-05 13:53:25 might be due to udev 2013-08-05 13:53:43 when i restart chronyd it sets time correctly 2013-08-05 13:56:31 ncopa, seems dmvpn is broke 2013-08-05 13:56:56 :-( 2013-08-05 13:57:14 fixable? 2013-08-05 13:57:45 yes, but not sure how fast. smells ip_gre bug. 2013-08-05 13:57:57 ok 2013-08-05 13:58:25 in any case, i will hold the v2.6.3 release til we got the grsecurity 3.10.5 patch 2013-08-05 13:58:32 oh 2013-08-05 13:58:42 maybe fix is in there 2013-08-05 13:58:55 there were like 100+ fixes from 3.10.4 to 3.10.5 2013-08-05 13:59:51 nope 2013-08-05 14:00:05 no ipv4 or gre ir similar 2013-08-05 14:00:21 ip_gre was rewritten for 3.10 2013-08-05 14:00:32 yeah 2013-08-05 14:00:38 i never had time to test it properly 2013-08-05 14:00:41 ok 2013-08-05 14:00:57 so there might be some work to get it fixed 2013-08-05 14:01:26 what do you think, we do v2.6.3 with the last 3.9 kernel? or we hold back v2.6.3 til we have gre fixed? 2013-08-05 14:05:30 ncopa, i'll take a look at it tomorrow 2013-08-05 14:12:06 fabled: thanks! 2013-08-05 21:12:38 n8@all 2013-08-06 04:17:33 Moinmoin 2013-08-06 06:19:03 fabled: grsec for 3.10.5 is out, I'm gonna upgrade the kernel 2013-08-06 06:20:00 ncopa, ok. i'm trying to figure out what's broke with dmvpn. 2013-08-06 07:29:26 fabled: did you look at it? are you close to a fix? otehriwse i just push 3.10.5 now 2013-08-06 07:30:42 ncopa, i've looked at it for few hours, but i am not sure yet. i think i need to do some printk() debugging to figure out where it breaks 2013-08-06 07:30:54 ok, I'll push it 2013-08-06 07:31:14 edge or -stable? 2013-08-06 07:31:18 edge 2013-08-06 07:31:20 only edge 2013-08-06 07:31:23 mmm. okay. 2013-08-06 07:31:37 so others have change to test other things meanwhile 2013-08-06 08:38:14 ncopa, ok i have one patch to test with ip_gre 2013-08-06 08:38:52 you want me to build an iso for you? 2013-08-06 08:38:57 or a /boot dir 2013-08-06 08:39:43 i think i should be able to test it 2013-08-06 08:39:47 though edge seems broke 2013-08-06 08:39:57 broke? 2013-08-06 08:40:01 libsasl upgraded, but spice-server not recompiled 2013-08-06 08:40:04 upgrade fails 2013-08-06 08:40:25 hm... i think i had that too 2013-08-06 08:40:33 but i thought i solved it... 2013-08-06 08:41:39 spice 0.12.4 should be newer that libsasl 2013-08-06 08:41:45 than* 2013-08-06 08:41:56 libsasl soname changed 2013-08-06 08:42:11 libsasl-2.1.23-r15: 2013-08-06 08:42:11 conflicts: libsasl-2.1.26-r2 2013-08-06 08:42:11 satisfies: spice-server-0.12.4-r0[so:libsasl2.so.2] 2013-08-06 08:43:12 http://sprunge.us/hPDA 2013-08-06 08:43:36 maybe i have a local build? 2013-08-06 08:43:56 ncopa-desktop:~$ sudo apk upgrade -a --simulate 2013-08-06 08:43:56 OK: 2525 MiB in 993 packages 2013-08-06 08:44:21 oh 2013-08-06 08:44:22 no 2013-08-06 08:44:27 i had edge + v2.6 2013-08-06 08:44:38 both untagged 2013-08-06 08:44:41 and it choce v2.6 version 2013-08-06 08:44:53 i wonder why 2013-08-06 08:45:09 something else picks it 2013-08-06 08:45:22 i think it locked it too early 2013-08-06 08:45:26 apk info -r so:libsasl2.so.3 2013-08-06 08:45:28 sounds minor apk bug 2013-08-06 08:45:30 apk info -r so:libsasl2.so.2 2013-08-06 08:46:19 its likely claws mail, subversion or something that needs libsasl2.so.2 2013-08-06 08:48:36 oh right... satisfies: spice-server-0.12.4-r0[so:libsasl2.so.2] 2013-08-06 08:48:38 oh yes, that's true it choce the v2.6 version of some higher item 2013-08-06 08:48:50 explains everything 2013-08-06 08:48:58 /usr/bin/fc-cache: symbol 'mkostemp64': can't resolve symbol in lib '/usr/lib/libfontconfig.so.1'. 2013-08-06 08:49:03 /usr/bin/fc-cache: symbol 'mkostemp64': can't resolve symbol in lib '/usr/lib/libfontconfig.so.1'. 2013-08-06 08:49:16 it needs newer uclibc... 2013-08-06 08:49:36 try apk upgrade --latest 2013-08-06 08:50:07 we need that symbol tracking to set versioned depends 2013-08-06 08:50:09 mmm 2013-08-06 08:50:53 i had local version of uclibc with same pkgrel 2013-08-06 08:51:07 okay 2013-08-06 08:51:10 seems it's good now 2013-08-06 09:11:29 algitbot: build master 2013-08-06 10:14:31 ncopa, i think my patch is working. will post formally soon. 2013-08-06 10:48:07 ncopa, http://patchwork.ozlabs.org/patch/264994/ 2013-08-06 11:00:34 man... you are fast... 2013-08-06 11:16:07 telehouse just racked new equipment without removing the packing materials... 2013-08-06 11:16:18 yeah, today's going to be an awesome day. i can already tell. 2013-08-06 11:17:39 kaniini, how on earth is that possible physically? .jpg or it didn't happen ;) 2013-08-06 11:18:28 fabled: on the supermicro microcloud stuff, they put cardboard down inside the chassis between the fans and the backplane to hold the backplane in place during shipping 2013-08-06 11:18:47 oh 2013-08-06 11:18:56 fabled: so i power up the chassis... all 8 blades throw PROCHOT immediately 2013-08-06 11:19:22 i see 2013-08-06 11:19:26 brilliant 2013-08-06 11:19:50 fabled: they put a warning label on the top of the chassis saying to remove the cardboard inside too... 2013-08-06 11:20:21 sounds like it's not the first or last time it happens... 2013-08-06 11:22:35 they probably did the same on the other chassis i had racked too 2013-08-06 11:22:46 haven't turned that one on yet. i told them to check both, hahah 2013-08-06 11:36:55 I have removed the packing materials from both of the chassis, we apologise for the mistake. 2013-08-06 11:36:58 Please verify the chassis and blades are not throwing any more alarms. 2013-08-06 11:37:01 OH YES. TODAY IS GOING TO BE A GREAT DAY. 2013-08-06 11:51:19 ... 2013-08-06 11:51:21 wow 2013-08-06 11:51:35 these guys plugged in eth1 to the external switch and eth0 to the internal 2013-08-06 11:51:38 thus breaking everything 2013-08-06 11:51:40 thanks guys! 2013-08-06 11:52:09 got to grep some popcorn... 2013-08-06 11:52:15 ouch 2013-08-06 11:54:03 i guess next time i should send them an MS Paint diagram 2013-08-06 11:54:11 or whatever we have in alpine that is like MS Paint 2013-08-06 11:57:07 my advice 2013-08-06 11:57:12 if you need to colocate in the uk 2013-08-06 11:57:19 get your own tech 2013-08-06 11:57:29 at least the ones at telehouse east are awful 2013-08-06 12:17:03 apk add mspain 2013-08-06 12:17:22 i feel with you kaniini 2013-08-06 12:17:49 there 2013-08-06 12:17:54 8 more alpine boxes installed 2013-08-06 12:17:56 :D 2013-08-06 13:51:13 im thinking that our initramfs init should have a tiny netlink listener 2013-08-06 13:51:23 then we trigger uevents 2013-08-06 13:51:55 the netlink listener will load any modaliases with libmount 2013-08-06 13:52:24 is the alpine_dev is found, then it will trigger the load of the rest 2013-08-06 13:52:49 i think we currently have an issue with mdev and usb keyboards 2013-08-06 13:53:12 issue #2071 2013-08-06 14:29:41 time gets wrong 2013-08-06 14:29:46 on bootup 2013-08-06 14:30:07 smells like hwclock --systohc UTC issue 2013-08-06 21:39:21 hm, about the keyboard 2013-08-06 21:39:29 I think that prymeroot had similiar issue 2013-08-06 21:41:18 that it was not detected? 2013-08-07 06:18:44 no usb keyb in initramfs emergency console? 2013-08-07 06:18:50 i think i found a solution for it 2013-08-07 06:24:50 fabled: btw, we need fix mkinitfs to handle musl 2013-08-07 06:24:53 i have some ideas 2013-08-07 06:25:10 yes 2013-08-07 06:25:17 currently we have a static list of files to be included in the initramfs 2013-08-07 06:25:25 uclibc included 2013-08-07 06:25:48 i have some ideas how to use ldd to automatically find all needed files 2013-08-07 06:26:15 so you can only say /bin/busybox and it will automatically pull in the linked so's 2013-08-07 06:27:27 right 2013-08-07 06:27:29 possible 2013-08-07 06:27:36 they fixed ldd in musl to report itself finally :) 2013-08-07 06:28:06 well, yes it will depend on a working ldd 2013-08-07 06:28:17 rebooting 2013-08-07 06:33:23 so, it was hid-generic driver that was missing in the initramfs 2013-08-07 06:33:27 and missing autoload of it 2013-08-07 06:33:50 i wonder if hid-generic should be in the mkinitfs 'usb' feature 2013-08-07 06:33:54 or in the 'base' feature 2013-08-07 06:34:07 i suppose it is theoritically possible to have like bluetooth keyb 2013-08-07 06:34:30 but I suppose for now the only user of hid-generic is usb 2013-08-07 07:17:44 i dont know if anyone noticed but it seems like util-linux's hwclock does something nasty with hwclock 2013-08-07 07:17:51 everytime i reboot the time is wrong 2013-08-07 07:18:16 i suspect this is the reason: https://github.com/karelzak/util-linux/commit/d53f8ecfbf036eddb2aef737dc0973a613d80ced 2013-08-07 07:29:37 ncopa, same here 2013-08-07 07:35:46 https://github.com/karelzak/util-linux/commit/910a090039cbd529041bfb5f6be72bf27a96bd47 2013-08-07 07:35:51 looks more suspicious 2013-08-07 07:51:20 no, this looks suspicious: https://github.com/karelzak/util-linux/commit/839be2ba6b44fa9dc927f081d547ebadec9de19c 2013-08-07 07:52:26 note that the if (!universal) tv.tv_sec += minutewets * 60; now only happens if (debug) 2013-08-07 08:12:36 yes, seems to be wrong 2013-08-07 08:13:05 i dunno 2013-08-07 08:14:03 makes my head hurt :) 2013-08-07 08:14:39 * The first call to settimeofday after boot will assume the systemtime 2013-08-07 08:14:39 * is in localtime, 2013-08-07 08:14:50 i dunno if that require some kernel config 2013-08-07 08:23:07 i wonder if the adtjtime thingy triggers it 2013-08-07 08:30:28 ncopa, did you have new edge snapshot .isos ? 2013-08-07 08:31:18 i'm on it 2013-08-07 08:33:56 https://github.com/karelzak/util-linux/blob/master/sys-utils/hwclock.c#L831 2013-08-07 08:34:04 it assumes that it is the first call 2013-08-07 08:34:21 however our init.d/hwclock does a hwclock call with --adtjtime 2013-08-07 08:34:37 http://dev.alpinelinux.org/~ncopa/alpine/alpine-edge-130807-x86.iso 2013-08-07 08:34:45 http://dev.alpinelinux.org/~ncopa/alpine/alpine-edge-130807-x86.iso.sha1 2013-08-07 08:36:41 in /etc/conf.d/hwclock: 2013-08-07 08:36:56 # If you want the hwclock script to set the system time (software clock) 2013-08-07 08:36:56 # to match the current hardware clock during bootup, leave this 2013-08-07 08:36:57 # commented out. 2013-08-07 08:36:57 # However, you can set this to "NO" ifyou are running a modern kernel 2013-08-07 08:36:57 # with CONFIG_RTC_HCTOSYS set to y and your hardware clock set to UTC. 2013-08-07 08:36:57 #clock_hctosys="YES" 2013-08-07 08:37:39 we have a such modern kernel 2013-08-07 08:37:52 we have CONFIG_RTC_HCTOSYS=y 2013-08-07 08:39:14 quad core via nano: http://www.picco.nl/catalog/product_info.php/manufacturers_id/13/products_id/4633/language/en 2013-08-07 08:39:23 i wonder if i should test out one 2013-08-07 08:39:56 anyone have experience? 2013-08-07 08:40:25 nope 2013-08-07 08:53:13 ncopa, have you seen http://xtables.de ? 2013-08-07 08:53:28 nope 2013-08-07 08:55:21 seems that http://netfilter.org/projects/nftables/ is again alive 2013-08-07 08:55:28 there's patchset for 3.10 in git 2013-08-07 08:55:45 nftables is the 'virtual machine' in kernel for packet filtering, and handling everything else in userspace 2013-08-07 08:57:31 nftables and xtables2 are 2 competing things? 2013-08-07 08:59:27 i will upgrade dahdi-linux to 2.7.0 2013-08-07 08:59:33 as 2.6.2 does not build against 3.10 kernel 2013-08-07 08:59:50 in 2.6-stable that is 2013-08-07 09:04:25 ncopa, good 2013-08-07 09:04:42 i think parts compete, parts not 2013-08-07 09:04:49 but yes, they are overlapping projects 2013-08-07 09:04:51 i'm off to lunch 2013-08-07 09:05:20 https://lwn.net/Articles/531752 2013-08-07 09:43:58 hi 2013-08-07 09:47:37 hi 2013-08-07 12:03:13 wiki.a.o is down :-( 2013-08-07 12:03:43 lists.a.o too 2013-08-07 12:03:48 probably same box 2013-08-07 12:04:03 what happened? 2013-08-07 12:04:08 dl-3.a.o 2013-08-07 12:04:10 i dunno 2013-08-07 12:04:45 ncopa-desktop:~$ ping wiki.alpinelinux.org 2013-08-07 12:04:45 PING wiki.alpinelinux.org (74.117.189.116) 56(84) bytes of data. 2013-08-07 12:04:45 From 216.55.14.66 icmp_seq=1 Destination Host Unreachable 2013-08-07 12:06:05 sounds like this is the datacenter change 2013-08-07 12:06:20 i think the sponsor for that box is changing datacenters 2013-08-07 12:07:33 18 * generic-con.car1.dallas1.level3.net (4.59.195.2) 1313.260 ms !H * 19 generic-con.car1.dallas1.level3.net (4.59.195.2) 1345.927 ms !H 3174.009 ms !H * 2013-08-07 12:08:28 yeah 2013-08-07 12:08:38 i thought it was on friday 2013-08-07 12:09:09 ok 2013-08-07 12:30:02 files from v2.5.4-177-g8631725 uploaded 2013-08-07 12:31:51 ncopa: last I heard the datacenter move was friday evening :/ 2013-08-07 12:32:16 maybe they powered it down til friday :) 2013-08-07 12:32:19 I have a backup box ready that I tested (unsuccessfully) yesterday to restore wiki.a.o on 2013-08-07 12:32:33 but I didn't grab the apkovl from mail.a.o yet 2013-08-07 12:32:36 and it's on dl-2 2013-08-07 12:32:47 ok 2013-08-07 12:32:56 lets hope we didnt lose them completely 2013-08-07 12:33:00 moving servers in friday evening 2013-08-07 12:33:05 this cannot end good 2013-08-07 12:33:08 :) 2013-08-07 12:33:09 ncopa: no, we get nightlies 2013-08-07 12:33:18 ncopa: and the nightlies get tested regularly 2013-08-07 12:33:27 it's the nice thing with using lbu as backup tool :) 2013-08-07 12:33:34 very easy to do test restores 2013-08-07 12:33:37 cool :) 2013-08-07 12:34:17 ncopa: there's 2 mirrors in that datacenter: dl-3,4 that are offline 2013-08-07 12:34:28 but we should be ok with the others that are online 2013-08-07 12:36:20 ncopa: i was just typing an email to alpine-devel ML to notify about the outage, but realizing that with mail.a.o offline that won't work very well lol 2013-08-07 12:36:25 ACTION needs coffee apparently 2013-08-07 12:36:26 lol 2013-08-07 12:36:28 :) 2013-08-07 12:37:48 jbilyk_: shout it out loud here and hope everybody hears it. :) 2013-08-07 12:37:58 I NEED COFFEE :) 2013-08-07 12:38:03 clandmeter: ^ 2013-08-07 12:38:06 lol 2013-08-07 12:38:10 about udev… I haven't decided yet if I'm going to sell my soul and build it from systemd and do required changes in initramfs or just package eudev 2013-08-07 12:38:13 i mean about the outage :) 2013-08-07 12:38:18 ncopa: ↑ 2013-08-07 12:38:30 barthalion: yeah, i was thinking of udev the other day 2013-08-07 12:38:42 i think i'd try the systemd's version first 2013-08-07 12:38:48 if that works, then fine 2013-08-07 12:38:51 if not.. 2013-08-07 12:38:54 NOTICE: dl-3/4, wiki, and mail.alpinelinux.org are currently offline because of a provider outage 2013-08-07 12:39:03 maybe it does, but require a lot of packaging job 2013-08-07 12:39:03 we're working on restoring service 2013-08-07 12:39:19 files from v2.4.11-74-g0b6330a uploaded 2013-08-07 12:39:21 jbilyk_: no colors? :) 2013-08-07 12:39:25 barthalion: complain upstream... :-p 2013-08-07 12:39:28 files from v2.4.11-74-g0b6330a uploaded 2013-08-07 12:39:31 yeah… 2013-08-07 12:39:34 clandmeter: it's morning, too complicated :P 2013-08-07 12:39:37 it's not as easy as Lennart stated 2013-08-07 12:39:58 "you guys promised this should be easy... now, please help!" 2013-08-07 12:40:14 and then, cat assasins will follow me 2013-08-07 12:40:20 and I doubt my dog will defend me 2013-08-07 12:40:36 yeah, they'll kill you for not using systemd 2013-08-07 12:41:42 are there any new features that we need in newer udev? 2013-08-07 12:42:10 nah, just one user on arch bbs complained that we have quite old udev 2013-08-07 12:42:49 i think newer udevs has mostly fixes for systemd 2013-08-07 12:42:57 but that said 2013-08-07 12:43:11 i would love an udev replacement project 2013-08-07 12:43:13 eudev merges udev related changes to its code 2013-08-07 12:43:16 like hotplug2 2013-08-07 12:43:23 I'll put it in testing soon 2013-08-07 12:43:26 ok 2013-08-07 12:43:29 sounds good 2013-08-07 12:43:44 i think writing somethign from scratch would be ok 2013-08-07 12:43:45 I guess that it's not ready yet for production usage 2013-08-07 12:43:48 shouldnt be too hard 2013-08-07 12:43:51 but maybe one day 2013-08-07 12:44:11 we need something that listens on netlink socket kuevents 2013-08-07 12:44:29 load kernel modules on MODALIAS 2013-08-07 12:44:36 and set permissions on /dev/* stuff 2013-08-07 12:44:41 thats it basically 2013-08-07 12:44:54 well, we need some libudev compat thing too 2013-08-07 12:45:12 i'm not sure but it looks like libudev is a frontend for /sys basically 2013-08-07 13:10:49 i'll look at it soon 2013-08-07 13:13:13 also been thinking of a simple version for initramfs 2013-08-07 13:13:23 a coldplugger 2013-08-07 13:13:44 it waits for a given block id/label or something 2013-08-07 13:13:50 where the apks are 2013-08-07 13:14:10 and when found it continues boot up with the found alpine_dev 2013-08-07 13:53:55 files from v2.3.6-257-ga3d144a uploaded 2013-08-07 13:54:09 files from v2.3.6-257-ga3d144a uploaded 2013-08-07 13:58:45 please git pull git://git.alpinelinux.org/fab/aports 2013-08-07 13:59:10 only some updates 2013-08-07 14:03:50 files from v2.5.4-178-g8839eda uploaded 2013-08-07 14:05:12 files from v2.3.6-258-ge45527d uploaded 2013-08-07 14:06:09 files from v2.3.6-258-ge45527d uploaded 2013-08-07 15:01:44 /sbin/mke2fs: symbol 'mkostemp': can't resolve symbol 2013-08-07 15:01:52 i assume this is already fixed in edge ? 2013-08-07 15:02:19 ALL: infra issues have been resolved - there was a routing issue in the datacenter hosting a few a.o boxes 2013-08-07 15:02:28 that's been cleared up 2013-08-07 15:15:43 yap 2013-08-07 15:15:44 it is 2013-08-08 08:50:53 files from v2.4.11-75-ga14b326 uploaded 2013-08-08 08:58:40 files from v2.4.11-76-gbae0972 uploaded 2013-08-08 08:58:41 files from v2.4.11-76-gbae0972 uploaded 2013-08-08 08:58:46 files from v2.5.4-179-g3015e5e uploaded 2013-08-08 08:58:53 files from v2.5.4-179-g3015e5e uploaded 2013-08-08 08:59:48 files from v2.3.6-259-g4b8d261 uploaded 2013-08-08 14:56:14 here comes firefox-23 :) 2013-08-08 14:56:17 and it actually works 2013-08-09 08:00:51 oh 2013-08-09 08:01:04 the apk del globbing is powerful 2013-08-09 08:01:11 apk del .makedepends-* 2013-08-09 08:01:16 wonderful! 2013-08-09 11:04:06 if there is any plan for php 5.5 sooner then testing/php-zend-opcache could be moved to unmaintained 2013-08-09 11:04:52 as 5.5 core includes Zend Opcache 2013-08-09 11:07:50 and breaks xcache and some other modules probably 2013-08-09 11:08:02 and it's not well supported by phpbb3 yet 2013-08-09 11:09:49 barthalion, you mean enable 2 cache same time? 2013-08-09 11:10:49 some people may want to stay on xcache, which doesn't build against php 5.5 2013-08-09 11:14:07 but moving to 5.5 is certain, it adds and enhances some features that would change most php programming patters ;) 2013-08-09 11:14:16 patterns 2013-08-09 11:15:54 sooner or later yes, but it doesn't mean we have to be one of the early adopters :p 2013-08-09 11:16:09 I am sure php-xcache people are working on it 2013-08-09 11:16:17 to make it 5.5 compat 2013-08-09 11:16:48 possibly 2013-08-09 11:16:58 but until then, at least I won't push 5.5 to repos 2013-08-09 11:17:33 though I still use apc for my local apps 2013-08-09 11:18:22 I think 5.5 could wait, atleast v2.7 2013-08-09 11:36:54 vkrishn: fire up your alpine-sdk and get us php 5.5 :) 2013-08-09 11:37:23 barthalion: which problems do you see with phpbb3? mysql? 2013-08-09 11:40:56 clandmeter, I could give a try, I still haven't mirrored v2.6 due to internet problems last few weeks 2013-08-09 11:41:16 but pls don't back on it 2013-08-09 11:41:58 2.6? for sdk you need edge repo. 2013-08-09 11:42:06 last few days internet company have worked on it and seems stable 2013-08-09 11:42:13 ok 2013-08-09 11:43:29 i dont see any issue for having 5.5 in edge. if i get some free time i do it myself. 2013-08-09 11:44:16 :) 2013-08-09 12:41:44 clandmeter: and breaks xcache and some other modules probably 2013-08-09 12:41:49 check that the modules builds 2013-08-10 16:43:28 it seems that xcache can be built easily 2013-08-10 16:43:35 but we will have to drop apc 2013-08-10 16:43:46 ncopa_, clandmeter ↑ 2013-08-10 16:44:21 I'll take a look tommorow 2013-08-10 16:44:32 and bump some other packages btw 2013-08-11 18:20:05 so, i'm trying to find the problem with quassel core 2013-08-11 18:20:22 first step: how to activate debug symbols? 2013-08-11 18:27:40 ah: 2013-08-11 18:27:42 -DCMAKE_BUILD_TYPE=Debug 2013-08-11 18:27:59 let's see what happens... 2013-08-11 19:46:19 hm, 2013-08-11 19:46:27 Reading symbols from /usr/bin/quasselcore...(no debugging symbols found)...done. 2013-08-11 19:46:33 with option 2013-08-11 19:46:42 -DCMAKE_BUILD_TYPE=Debug 2013-08-11 19:46:47 any ideas why? 2013-08-12 02:33:47 StarWarsFan|afk: build it with features="!strip" 2013-08-12 06:12:23 ncopa, clandmeter: xcache from svn builds fine against php 5.5 2013-08-12 06:12:48 when will they make release? 2013-08-12 06:14:42 xcache 3.1 2013-08-12 06:15:18 https://groups.google.com/forum/#!topic/xcache/KPcc_duWcSI 2013-08-12 06:15:42 claims he wants some real-world apps in php 5.5 tested 2013-08-12 06:16:11 i believe it would be wise to wait for xcache-3.1 2013-08-12 06:16:32 but if ppl are in a hurry with testing php-5.5, then i suppose we could backport a patch 2013-08-12 06:16:51 optjonally, we could do php-5.5 in testing/ 2013-08-12 06:19:05 or just wait as I said 2013-08-12 06:19:12 we will have to drop apc anyway 2013-08-12 06:19:52 thanks for updating vim and the others :) 2013-08-12 06:20:00 np :) 2013-08-12 06:20:16 so let's wait for stable xcache 3.1 2013-08-12 06:20:24 yeah, sounds good to me 2013-08-12 06:20:59 i am rebuilding world 2013-08-12 06:21:13 to catch bison-3 breakgages 2013-08-12 06:21:24 and new abuild 2013-08-12 06:21:41 and the recent --host --build change in almost all apkbuilds 2013-08-12 06:51:08 moin 2013-08-12 06:51:18 ncopa: did you read backlog? 2013-08-12 06:51:26 yesterday starting 20:20 2013-08-12 06:52:11 current state: i build quassel directly on the cmdline, now i got debug symbols 2013-08-12 06:52:23 but gdb still shows an empty backtrace 2013-08-12 06:52:25 any ideas? 2013-08-12 06:56:23 kaniini: ah, thx for the hint. will try this too... 2013-08-12 07:16:09 I want to have collaboration with you 2013-08-12 07:20:14 I am a master of scientific computing . 2013-08-12 07:33:38 StarWarsFan: no i lost backlog. power went off yesterday due to lightning 2013-08-12 07:33:46 will check the irc logger log 2013-08-12 07:33:52 snima: hi, welcome 2013-08-12 07:34:40 snima: http://wiki.alpinelinux.org/wiki/Contribute 2013-08-12 07:35:31 have you ever received a postcard? sounds cool 2013-08-12 07:38:40 lol 2013-08-12 07:38:43 no :) 2013-08-12 07:39:38 what a pity :p 2013-08-12 07:39:52 i'm pretty ok with it :) 2013-08-12 07:40:38 maybe i should make a wish list though... 2013-08-12 08:13:01 ncopa: i made you an excerpt: 2013-08-12 08:13:03 [20:20:05] so, i'm trying to find the problem with quassel core 2013-08-12 08:13:03 [20:20:22] first step: how to activate debug symbols? 2013-08-12 08:13:03 [20:27:40] ah: 2013-08-12 08:13:03 [20:27:42] -DCMAKE_BUILD_TYPE=Debug 2013-08-12 08:13:03 [20:27:58] let's see what happens... 2013-08-12 08:13:05 [21:46:19] hm, 2013-08-12 08:13:07 [21:46:26] Reading symbols from /usr/bin/quasselcore...(no debugging symbols found)...done. 2013-08-12 08:13:10 [21:46:33] with option 2013-08-12 08:13:12 [21:46:42] -DCMAKE_BUILD_TYPE=Debug 2013-08-12 08:13:14 [21:46:47] any ideas why 2013-08-12 08:13:16 after that I build by hand using this cmd: 2013-08-12 08:13:18 cmake -DCMAKE_INSTALL_PREFIX=/usr/ -DWITH_KDE=OFF -DCMAKE_BUILD_TYPE=RelWithDebInfo -DWITH_OPENSSL=ON -DWANT_CORE=ON -DWANT_MONO=ON -DWANT_QTCLIENT=ON -DWITH_DBUS=OFF -DWITH_OXYGEN=OFF -DWITH_PHONON=OFF -DWITH_WEBKIT=OFF -DLINGUAS="de en_US" ../quassel-0.9.0/ 2013-08-12 08:13:21 and started quassel this way with gdb: 2013-08-12 08:13:23 gdb /tmp/quasselcore -- --configdir=/var/lib/quassel --debug --logfile=/var/log/quassel.log --loglevel=debug 2013-08-12 08:13:25 ... 2013-08-12 08:13:27 Reading symbols from /tmp/quasselcore...done. 2013-08-12 08:13:29 ... 2013-08-12 08:13:32 but as soon a quassel client connects, quasselcore crashes but the backtracke is empty!? 2013-08-12 08:32:47 ncopa: thx :) 2013-08-12 08:33:13 ncopa: I am looking for a scientific work 2013-08-12 08:47:14 StarWarsFan: a bit busy here now will check it up later 2013-08-12 08:49:24 ok 2013-08-12 13:46:18 Can I request mkdir -p /var/run/xen be created by default? 2013-08-12 14:20:04 Frosh: yes 2013-08-12 14:20:13 i thought i had fixed that already 2013-08-12 15:06:52 StarWarsFan: sorry been horribly busy today 2013-08-12 15:07:12 add the -DCMAKE_BUILD_TYPE=Debug and then build it with: DEBUG=1 abuild -r 2013-08-12 15:07:22 otherwise will abuild strip the symbols 2013-08-12 15:07:39 optionally i think you can add to options="!strip" in APKBUILD 2013-08-12 15:10:24 Frosh: in case i forget it, could you please craete an issue on bugs.a.o about that mkdir -p /var/run/xen 2013-08-12 15:10:33 thanks 2013-08-12 15:11:29 ah 2013-08-12 15:11:31 its there already 2013-08-12 15:11:33 http://bugs.alpinelinux.org/issues/2031 2013-08-12 15:13:13 thanks ncopa 2013-08-12 15:13:38 sorry that i missed that for v2.6.3 2013-08-12 15:13:50 i'll fix it asap 2013-08-12 15:13:53 hopefully tm 2013-08-12 15:13:57 I like how alpine works out the box on old mac mini 2013-08-12 15:14:11 i dunno 2013-08-12 15:14:19 gotta run now 2013-08-12 15:14:21 see you 2013-08-12 15:14:25 see ya 2013-08-12 18:27:25 is there a package with libc debug symbols? 2013-08-12 18:27:47 can't find anything like *-dbg 2013-08-12 18:28:01 except kamailio-dbg 2013-08-12 18:30:13 I doubt it 2013-08-12 18:30:24 you're on your own 2013-08-12 18:38:00 great... :-( 2013-08-12 18:40:07 so i have to build a special version of libc package? 2013-08-12 18:40:12 with debug symbols? 2013-08-12 18:40:19 exactly 2013-08-12 18:43:08 should i increase pkgrel for this? 2013-08-12 18:44:45 i did... 2013-08-12 18:44:48 building... 2013-08-12 18:47:55 i'm going to add a dbg() handler 2013-08-12 18:48:06 to abuild 2013-08-12 18:50:25 hm, build failed 2013-08-12 18:50:40 if building this way: 2013-08-12 18:50:55 .../main/libc0.9.32$ DEBUG=1 abuild -r 2013-08-12 18:51:04 kaniini: ah cool! 2013-08-12 18:51:30 ACTION is still on the problem with crashing quasselcore binary 2013-08-12 18:52:18 but there is no backtrace :-( 2013-08-12 18:55:55 ulimit -c unlimited 2013-08-12 18:56:03 gdb quasselcore /path/to/core 2013-08-12 18:56:09 the core will be dumped in cwd 2013-08-12 18:57:39 ok, thx, trying this 2013-08-12 18:59:49 ok, now i got a 40M core file 2013-08-12 19:06:51 here's the creation of the dump: 2013-08-12 19:06:53 http://pastebin.com/TJzz5dhf 2013-08-12 19:07:51 and this the result: 2013-08-12 19:07:52 http://pastebin.com/actBxHG9 2013-08-12 19:08:53 but: 2013-08-12 19:08:54 warning: core file may not match specified executable file. 2013-08-12 19:08:57 hm... 2013-08-12 20:33:06 hi 2013-08-12 20:33:18 'main/perl-io-socket-ssl: upgrade to 1.953' in aports broke git send-email 2013-08-12 20:33:26 in 1.950 the default of ssl_verify_mode was changed from SSL_VERIFY_NONE to SSL_VERIFY_PEER in perl-io-socket-ssl 2013-08-12 20:33:34 ca-certificates should be added as a dep for either perl-io-socket-ssl or git-perl to fix it 2013-08-12 20:34:19 since git-perl seems to be the only reverse depends I think it will be best to add it there 2013-08-12 20:35:02 as other consumers of perl-io-socket-ssl could override the default ssl_verify_mode setting 2013-08-12 20:35:55 and therefore make the ca-certificates dep unnecessary 2013-08-12 20:46:46 related arch linux bug: https://bugs.archlinux.org/task/36506 2013-08-13 04:33:21 Morning 2013-08-13 05:04:48 morning 2013-08-13 05:10:18 morning 2013-08-13 05:15:20 crazy hour to wake up 2013-08-13 05:47:22 Depends on the timezone ;-) 2013-08-13 05:48:58 right, I assumed you're in CET :) 2013-08-13 06:16:15 morning 2013-08-13 06:26:24 uggedal: umm, yeah, I've seen your comment on Arch bugtracker 2013-08-13 06:26:48 ncopa: https://bugs.archlinux.org/task/36506 2013-08-13 06:26:58 ncopa: do you want to fix it or I can do it? 2013-08-13 06:27:28 not sure where I should add ca-certificates as dependency, probably to perl-io-socket-ssl due to current default behaviour 2013-08-13 06:28:22 hm 2013-08-13 06:28:32 i think that makes sense yes 2013-08-13 06:28:55 barthalion: feel free to fix it 2013-08-13 06:28:56 thanks 2013-08-13 06:29:01 np :) 2013-08-13 06:29:27 and thanks for asking :) 2013-08-13 06:33:17 uggedal: fixed :) 2013-08-13 06:42:25 barthalion: switzerland :-) 2013-08-13 10:53:07 hi 2013-08-13 10:55:53 hi 2013-08-13 11:25:33 incredible, libvirt isleaking 8.3MB/sec when i connect virt-manager to it 2013-08-13 11:34:02 latest release? 2013-08-13 15:20:18 btw, did not made some progress with the quassel problem yesterday :-( 2013-08-13 15:20:40 kaniini: what do you think when abuild is able to build dbg-pkg's? 2013-08-13 15:23:52 soon 2013-08-13 15:23:54 i have like 2013-08-13 15:23:57 gigantic todo list 2013-08-13 15:23:59 lol 2013-08-13 15:24:27 ah, so welcome to the club ;-) 2013-08-13 15:38:18 do you get paid for this work? 2013-08-13 16:09:40 shafire: yes 2013-08-13 17:57:05 kaniini: from whom? 2013-08-13 17:58:42 ncopa: I'm going to push ffmpeg 2.0 soon 2013-08-13 17:59:00 ncopa: It took me a while to patch everything in Arch, but obviously it should be faster here :) 2013-08-13 17:59:47 kaniini: who hasn't such list… 2013-08-13 17:59:54 shafire: you can pay me, I won't be mad 2013-08-13 18:00:23 whenever I will be a millionaire, I will do it 2013-08-13 18:06:55 ncopa: http://sprunge.us/fPhC I'm actually wondering… shouldn't linux-headers match the lowest pkgver from kernels we provide? 2013-08-13 18:09:48 shafire: we use alpine exclusively for our infrastructure 2013-08-13 18:09:57 shafire: so, working on alpine is a component of my day job 2013-08-13 18:10:28 shafire: i would figure most of the core team is being paid, at least, indirectly, to work on alpine some way or another 2013-08-13 18:27:45 clandmeter: how do you generate kernel config for both architectures? is there something more clever than two different chroots? 2013-08-13 18:29:26 oops, wrong person 2013-08-13 18:29:30 rnalrd: ↑ 2013-08-13 19:04:57 wat 2013-08-13 21:21:22 n8@all 2013-08-14 04:31:08 Moinmoin 2013-08-14 06:19:52 morning 2013-08-14 06:22:09 barthalion: i think you have a point with kernel headers 2013-08-14 06:22:18 but i dont think it matters that much 2013-08-14 06:23:06 linux-scst should be removed 2013-08-14 06:23:23 linux-vserver can now be upgraded to 3.10.y 2013-08-14 06:23:47 so vanilla and virt-grsec 2013-08-14 06:24:58 yeah 2013-08-14 06:25:35 oh and the linu-grsec kernel should probably have its config tweaked a bit too 2013-08-14 06:25:42 there was some features i'd like to enable 2013-08-14 06:31:40 oh well, I haven't spotted it's samba from testing… 2013-08-14 07:46:00 ncopa: I think because of the recent changes, the ccache symlinks dont work anymore. 2013-08-14 07:47:45 recent changes? 2013-08-14 07:48:17 it worked before :) 2013-08-14 07:48:41 [all autotools packages]: normalize ./configure maybe? 2013-08-14 07:49:42 checking for x86_64-alpine-linux-uclibc-cc... x86_64-alpine-linux-uclibc-cc 2013-08-14 07:49:45 checking whether the C compiler works... no 2013-08-14 07:56:34 ncopa: any idea? 2013-08-14 07:57:32 have you set CC? 2013-08-14 07:57:41 which version of abuild is it? 2013-08-14 07:57:42 abuild -v 2013-08-14 07:57:52 no... wait 2013-08-14 07:57:57 apk version abuild 2013-08-14 07:58:11 abuild-2.16.0-r0 = 2.16.0-r0 2013-08-14 07:58:26 grep CC /etc/abuild.conf 2013-08-14 07:58:36 nada 2013-08-14 07:58:48 hmmmm 2013-08-14 07:59:12 i use export PATH="/usr/lib/ccache/bin/:$PATH" 2013-08-14 07:59:22 this should have fixed it: http://git.alpinelinux.org/cgit/abuild/commit/?id=11d9a7d6018fd13587259edd1030a608e9f85449 2013-08-14 07:59:37 try remove ccache symlinks 2013-08-14 07:59:51 and set 2013-08-14 08:00:00 export CC="ccache gcc" 2013-08-14 08:00:00 export CXX="ccache g++" 2013-08-14 08:00:05 in /etc/abuild.conf 2013-08-14 08:00:14 so thats the prefered way to use ccache? 2013-08-14 08:00:20 thats how i do it atleast 2013-08-14 08:00:41 i see /usr/lib/ccache/bin/ used in other dists. 2013-08-14 08:00:49 thats also a possibility 2013-08-14 08:01:30 but the error you saw happens when there are no CC set i think 2013-08-14 08:01:37 what happens if you set CC=gcc? 2013-08-14 08:01:46 with that path 2013-08-14 08:01:57 and prefer ccache symlink 2013-08-14 08:02:50 clandmeter: can you in the APKBUILD that breaks add an msg "DEBUG: CC=$CC" 2013-08-14 08:02:59 ok 2013-08-14 08:03:06 and then tell me if CC gets set 2013-08-14 08:04:10 >>> php: DEBUG: CC=gcc 2013-08-14 08:05:08 ok.. weird 2013-08-14 08:05:14 it should have worked then 2013-08-14 08:08:37 /usr/lib/ccache/bin/x86_64-alpine-linux-uclibc-cc 2013-08-14 08:08:39 ccache: FATAL: Could not find compiler "x86_64-alpine-linux-uclibc-cc" in PATH 2013-08-14 08:09:40 what happens if you neutralize the PATH, and do the export CC="ccache gcc" way? 2013-08-14 08:11:33 same thing 2013-08-14 08:11:34 configure:3942: x86_64-alpine-linux-uclibc-cc -Os -fomit-frame-pointer -Os -fomit-frame-pointer -Wl,--as-needed conftest.c >&5 2013-08-14 08:11:37 ccache: FATAL: Could not find compiler "x86_64-alpine-linux-uclibc-cc" in PATH 2013-08-14 08:12:09 sounds like it is not respecting CC 2013-08-14 08:12:45 i am rebuilding world here 2013-08-14 08:12:55 so i'll bump into it eventually 2013-08-14 08:14:08 i suppose i should try build php 2013-08-14 08:14:44 building now... 2013-08-14 08:14:52 weird 2013-08-14 08:14:55 it works here 2013-08-14 08:14:58 on x86 2013-08-14 08:17:30 testing x86_64 2013-08-14 08:17:49 works too 2013-08-14 08:17:54 is it php-5.4? 2013-08-14 08:17:56 or 5.5 2013-08-14 08:18:37 5.5 2013-08-14 08:18:44 but i have same issue on 5.4 2013-08-14 08:18:53 works like a charm here... 2013-08-14 08:19:08 i guess its because i have: PATH="/usr/lib/ccache/bin/:$PATH" 2013-08-14 08:19:18 if you remove that? 2013-08-14 08:19:22 does it work then? 2013-08-14 08:19:28 I think so 2013-08-14 08:19:35 but then I have no ccache 2013-08-14 08:19:41 except if i export it directly 2013-08-14 08:20:31 hm 2013-08-14 08:20:36 ok 2013-08-14 08:20:38 i have an idea 2013-08-14 08:20:48 try rebuild ccache 2013-08-14 08:20:55 might be its the --host --build change 2013-08-14 08:21:04 that makes a difference 2013-08-14 08:21:15 i did 2013-08-14 08:22:23 im already on r1 2013-08-14 08:22:38 ok 2013-08-14 08:22:46 then the only suggestion i have 2013-08-14 08:22:47 just regular exporting CC and CCX works ok. 2013-08-14 08:22:54 is ln -s 2013-08-14 08:24:12 ln -s gcc /usr/bin/x86_64-alpine-linux-uclibc-cc 2013-08-14 08:24:26 we might need to do that in gcc package 2013-08-14 08:25:38 would be nice if we have a unified way to enable ccache. and we can add it to wiki. 2013-08-14 08:26:29 I'd suggest export CC="ccache gcc" 2013-08-14 08:43:50 ok 2013-08-14 20:33:47 n8@all 2013-08-14 20:54:06 ich 2013-08-14 21:04:27 du 2013-08-14 21:04:55 er/sie/es 2013-08-14 21:09:24 German is fun until you have to learn it 2013-08-14 21:09:55 Why? 2013-08-14 21:10:28 bah, just didn't like it 2013-08-14 21:11:03 you know, makeup exam and so on :p 2013-08-14 21:12:39 Heh, what can I say, I am forced to learn it :p 2013-08-14 21:13:59 fortunately I wasted only one holidays to prepare for exam 2013-08-14 21:14:54 What is a makeup exam? 2013-08-14 21:15:37 hm, maybe I translated it incorrectly 2013-08-14 21:16:11 an exam you write in the end of holidays if you failed on some subject 2013-08-14 21:16:52 what's your native language? 2013-08-14 21:17:02 Polish 2013-08-14 21:17:16 my surname is polish :) 2013-08-14 21:17:20 haha :D 2013-08-14 21:17:55 now I think I just didn't want to learn German and I regret that a bit 2013-08-14 21:20:15 shafire: are you from Germany? :) 2013-08-14 21:20:20 yes 2013-08-14 21:21:58 if you live close to Berlin or Szczecin, one day we can go for a beer :p 2013-08-14 21:22:21 No, I am from Stuttgart :-) 2013-08-14 21:23:00 ah 2013-08-14 21:23:01 you are living at the border 2013-08-14 21:23:47 in Szczecin, not far from border anyway 2013-08-14 21:23:52 to be bilingual is advantageous 2013-08-14 21:24:03 I'm actually moving to Poznań this month 2013-08-14 21:24:42 sure, especially when you live literally in the neighbourhood of country 2013-08-14 21:25:15 I guess that more than half of job offers here require German language 2013-08-14 21:31:57 good night 2013-08-14 21:34:47 good night 2013-08-14 21:34:48 see you! 2013-08-15 04:22:26 Moinmoin 2013-08-15 04:40:27 ncopa: kaniini: another one on the quassel irc channel compiled quassel on alpine 2013-08-15 04:40:31 with the same result 2013-08-15 04:40:50 quasselcore crashes if a client connects 2013-08-15 04:41:08 they assume that there is something wrong with qt 2013-08-15 04:41:50 unfortunately i have no idea how dig deeper into this 2013-08-15 04:41:59 no c experience 2013-08-15 04:42:05 ACTION is java developer ;-) 2013-08-15 04:42:40 so how to go ahead now? 2013-08-15 05:59:18 morrn 2013-08-15 05:59:46 StarWarsFan: do you have it built with debugging symbols? 2013-08-15 06:00:00 do you have a backtrace? 2013-08-15 06:16:15 bah, ada = bootstrap problem 2013-08-15 06:16:24 you need ada in order to build ada 2013-08-15 06:19:13 ncopa: the backtrace is the problem: it is always empty 2013-08-15 06:19:31 but yes, i have a debug build 2013-08-15 06:19:36 build as you said some days ago 2013-08-15 06:19:46 ok 2013-08-15 06:19:56 then it probably happens in other lib 2013-08-15 06:20:08 as they said on the quassel channel 2013-08-15 06:20:27 dont you have any symbol at all in the backtrace? 2013-08-15 06:20:49 the one and only thing i got is the following out of the quassel error log: 2013-08-15 06:21:16 Quassel IRC: 0.9.0 3b9dd1077c243d15258913fe0d34d53d082b8018 2013-08-15 06:21:16 # 0 quasselcore 0x00000a21910e3291 Quassel::logBacktrace(QString const&) 2013-08-15 06:21:16 # 1 quasselcore 0x00000a21910c9015 Quassel::handleSignal(int) 2013-08-15 06:21:16 # 2 libc.so.0.9.32 0x000073a677d3dc74 0x0000000000000000 2013-08-15 06:21:16 # 3 libc.so.0.9.32 0x0000000000000000 0x0000000000000000 2013-08-15 06:21:50 but if using gdb, the output is like this: 2013-08-15 06:21:51 ... 2013-08-15 06:21:53 2013-08-12 18:49:08 Info: Client ::FFFF:C0A8:2A40%0 initialized and authenticated successfully as "starwarsfan" (UserId: 1). 2013-08-15 06:21:53 [New LWP 1977] 2013-08-15 06:21:53 Program received signal SIGSEGV, Segmentation fault. 2013-08-15 06:21:53 [Switching to LWP 1977] 2013-08-15 06:21:53 0x00007136eeae2c4e in ?? () 2013-08-15 06:21:54 (gdb) bt 2013-08-15 06:21:57 #0 0x00007136eeae2c4e in ?? () 2013-08-15 06:21:59 #1 0x00007136ef8c1348 in ?? () 2013-08-15 06:22:01 #2 0x00007136ef9bd478 in ?? () 2013-08-15 06:22:21 and thats it? 2013-08-15 06:22:27 and this goes up to #37 2013-08-15 06:22:34 right, thats it 2013-08-15 06:22:48 Client ::FFFF:C0A8:2A40%0 2013-08-15 06:22:56 does same thing happen with ipv4? 2013-08-15 06:25:24 i can start quassel via gdb with "run" 2013-08-15 06:25:24 but as soon as i connect with the client, it crashes with the seg fault 2013-08-15 06:26:44 ipv4? i have no special setting regarding ipv4 as the core is running on a vmware virtual machine and the client is running on the host system 2013-08-15 06:27:46 my setup: 2013-08-15 06:28:03 host: win7-64 with vmare-workstation and quassel-client 2013-08-15 06:28:19 vm: alpine (edge-64) 2013-08-15 06:29:00 brb, have to leave the train and then some meetings... 2013-08-15 06:29:01 brb 2013-08-15 06:29:22 and read the backlog of course :-) 2013-08-15 08:13:12 re 2013-08-15 09:16:12 please git pull git://git.alpinelinux.org/fab/aports 2013-08-15 09:16:47 does it build? 2013-08-15 09:19:23 it seems like it should 2013-08-15 09:20:12 I'll push it in a moment 2013-08-15 09:29:54 hi 2013-08-15 09:35:14 since my last request to pull some time passed...there is always the possibility that my stuff breaks 2013-08-15 09:36:00 barthalion: i'm available for the next couple of hours to fix my thing 2013-08-15 09:36:56 I just screwed my local clone, so it may take a while 2013-08-15 09:37:26 fabian_a: if you could rebase your changes against aports tree next time it would be great 2013-08-15 09:40:04 what does a rebase do? 2013-08-15 09:41:12 barthalion: i always do that but this time it's a week ago. i really dislike to force my pushes because this can make things worse but there seems no other way 2013-08-15 09:41:15 moves your changes on top of repository 2013-08-15 09:41:22 shafire: ↑ 2013-08-15 09:42:32 hm, it went fine know 2013-08-15 09:42:38 barthalion: please check again 2013-08-15 09:42:45 ah, that's why :D 2013-08-15 09:43:32 fabian_a: thanks 2013-08-15 09:45:40 hmmm, mr failed 2013-08-15 09:46:21 it has been renamed to myrepos 2013-08-15 09:46:41 hm, but it's not the case 2013-08-15 09:47:04 there is no 1.15 tarball on server 2013-08-15 09:47:09 just 1.16 2013-08-15 09:49:53 I'll fix it :) 2013-08-15 09:50:40 Sigh, something is wrong either with my disk or repo 2013-08-15 10:00:30 sigh 2013-08-15 10:00:49 yeah, it's metapackage now 2013-08-15 10:00:57 so it is the case 2013-08-15 10:01:05 the issue is that this package not longer available. the tarball hosted at debian is empty 2013-08-15 10:01:55 fabian_a: I'll revert the changes; try to package myrepos 2013-08-15 10:02:13 renaming the package to myrepos and updating to 1.20130710 did the trick 2013-08-15 10:04:38 I wonder if we have kind of replace= in abuild 2013-08-15 10:05:22 yup 2013-08-15 10:10:15 argh 2013-08-15 10:13:22 algitbot: better build it 2013-08-15 10:13:34 algitbot: I'm a werewolf 2013-08-15 10:15:08 good for you 2013-08-15 10:23:58 barthalion: i changed the url and the source which now points to the upstream location instead of debian 2013-08-15 10:24:42 barthalion: it would be nice if you try to pull again from my repo 2013-08-15 10:26:50 barthalion: i'm not sure if my change will break myrepos/mr again because it seems that we changed the same stuff in different commit 2013-08-15 10:27:17 barthalion: rebase went fine 2013-08-15 10:35:37 sure, will do 2013-08-15 10:35:49 I have to finish some work first :P 2013-08-15 10:36:40 thanks 2013-08-15 12:10:40 kaniini: are you ok if i upgrade xen to 4.3.0 in edge? 2013-08-15 14:43:52 ncopa: ping 2013-08-15 14:48:07 barthalion: pong 2013-08-15 14:49:19 ncopa: what's the fix for issues with new bison? 2013-08-15 14:49:57 depends on the program or is there something universal? 2013-08-15 14:50:43 depends on the prog 2013-08-15 14:51:09 so far i have been able to find upstream patches for the things i have bumped into 2013-08-15 15:07:08 MySQL will probably fail 2013-08-15 15:07:14 at least MariaDB does 2013-08-15 15:07:32 I'm writing a bug report to the latter now 2013-08-15 19:19:10 ncopa: yes. 2013-08-15 19:19:18 ncopa: just be sure to rebase the patches. 2013-08-16 02:55:41 i'm going to update my host, so dl-5 will be down for a few minutes 2013-08-16 04:09:42 dl-5 should be up and running again 2013-08-16 07:24:13 qemu-1.6 is out 2013-08-16 08:07:34 moinmoin 2013-08-16 08:12:17 fabled: can you please help me with this: 2013-08-16 08:12:19 http://git.alpinelinux.org/cgit/aports/commit/main/openssh?id=5c904717322f8113870abf48cc93c3c8b00ab4e8 2013-08-16 08:12:24 openssh no longer builds 2013-08-16 08:12:54 the openssh-fix-includes.diff and openssh-fix-utmp.diff appears to be missing 2013-08-16 08:15:58 urgh 2013-08-16 08:16:00 sorry 2013-08-16 08:16:07 musl related 2013-08-16 08:16:11 you can remove them 2013-08-16 08:16:44 i'll add them later. or you think i should add them now? 2013-08-16 08:16:54 i dunno 2013-08-16 08:17:07 remove for now 2013-08-16 08:17:12 ok 2013-08-16 08:18:55 building qemu-1.6 now 2013-08-16 08:59:52 algitbot: build master 2013-08-16 10:43:19 ncopa, do we have https://github.com/moteus/lua-odbc/ ? 2013-08-16 10:44:21 or anything to access odbc or tds databases in lua 2013-08-16 11:05:16 i dunno 2013-08-16 11:05:29 do you want me build a pkg for lua-odbc? 2013-08-16 11:25:08 ncopa, nevermind. i think i can use asterisk-odbc instead 2013-08-16 11:25:39 it likely even works better for me 2013-08-16 11:25:59 oh, nice 2013-08-16 12:31:16 barthalion: do you know of a fix for wine 1.6 and bison-3? 2013-08-16 12:31:24 same as this: https://bugs.gentoo.org/show_bug.cgi?id=478672 2013-08-16 12:34:24 found it 2013-08-16 12:34:28 https://github.com/mirrors/wine/commit/8fcac3b2bb8ce4cdbcffc126df779bf1be168882 2013-08-16 12:37:53 ACTION is leaving into the weekend 2013-08-16 12:37:54 c ya 2013-08-16 13:43:20 ncopa, libusbx seems to conflict with libusbx ? 2013-08-16 13:43:35 ERROR: libusbx-1.0.16-r0: trying to overwrite usr/lib/libusb-1.0.so.0.1.0 owned by libusb-1.0.9-r0. 2013-08-16 13:43:35 ERROR: libusbx-1.0.16-r0: trying to overwrite usr/lib/libusb-1.0.so.0 owned by libusb-1.0.9-r0. 2013-08-16 13:43:43 similar with -dev pkg 2013-08-16 13:45:32 yes 2013-08-16 13:45:45 i removed libusb in edge 2013-08-16 13:46:06 i suppose i should add 'replaces=libusb' 2013-08-16 13:46:07 libusbx should conflict 2013-08-16 13:46:28 i have libusb-dev here in world 2013-08-16 13:46:41 replace it with libusbx-dev 2013-08-16 13:46:45 yes 2013-08-16 13:46:47 doing it 2013-08-16 13:47:02 but apk should have told me that libusb and libusbx conflict 2013-08-16 13:47:11 ah 2013-08-16 13:47:20 i should have added a depend=!libusb 2013-08-16 13:47:23 yes 2013-08-16 13:47:57 i was lazy and expected the file conflict handling deal with that ;) 2013-08-16 13:48:14 actually 2013-08-16 13:48:21 i think we have a circular dep there now 2013-08-16 13:48:27 libusbx depends on udev 2013-08-16 13:48:41 and iirc udev depends on libusb-compat 2013-08-16 13:48:45 which depends on libusbx 2013-08-16 13:49:05 we might need look into upgrade udev 2013-08-16 13:49:45 hm 2013-08-16 13:49:55 i think libusbx might be abi compat with libusb 2013-08-16 13:50:03 shoudl maybe been replaces=libusb 2013-08-16 13:50:13 rather than depend=libusb 2013-08-16 13:54:23 fabled: whats this: https://dpaste.de/Ww879/ 2013-08-16 13:54:37 i wonder why libcue is not upgraded 2013-08-16 13:54:47 or libva-intel-driver 2013-08-16 13:55:00 try 'apk policy ' 2013-08-16 13:55:39 https://dpaste.de/aZqGq/ 2013-08-16 13:55:53 that's weird 2013-08-16 13:56:18 similar with libva-intel-driver 2013-08-16 13:56:23 something makes it ignored. perhaps arch is wrong? 2013-08-16 13:56:49 but sounds really like a bug 2013-08-16 13:57:23 sounds like a bug yes 2013-08-16 13:57:41 i doubt arch is wrong 2013-08-16 13:57:56 it could be something related install_if 2013-08-16 13:58:10 quite possible 2013-08-16 13:58:26 perhaps the newer versions' install_if is not true 2013-08-16 14:03:51 libcue-1.4.0-r1 is required by: 2013-08-16 14:03:51 audacious-plugins-3.3.4-r1 2013-08-16 14:04:02 which was pulled in by install_if 2013-08-16 14:18:18 ah 2013-08-16 14:18:30 i got bitten of same libusb error 2013-08-16 14:18:33 i think 2013-08-16 14:18:51 ncopa-desktop:/var/lib/libvirt$ sudo apk del libusb 2013-08-16 14:18:51 World updated, but the following packages are not removed due to: 2013-08-16 14:18:51 libusb: upower libopenobex obex-data-server gnome-bluetooth spice-glib 2013-08-16 14:18:51 spice-gtk virt-manager spice-gtk-tools 2013-08-16 14:33:59 algitbot: build master 2013-08-17 04:14:05 Does anyone know of a working rust compiler for alpine? 2013-08-17 04:24:29 Ah, okay. I would like to put a request in for a rust compiler package for Alpine 2013-08-17 04:24:54 I would test compiling and installing this locally, but none of my alpine machines are specced high enough to do so 2013-08-17 04:25:03 Unless a chroot would suffice 2013-08-17 04:54:47 Hmm 2013-08-17 04:54:53 I tried to build rust from a chroot 2013-08-17 21:13:19 I can take a look soon 2013-08-17 21:13:22 Heartmender: ↑ 2013-08-18 03:28:59 barthalion: Sorry, was AFP 2013-08-18 03:29:01 AFK* 2013-08-18 03:29:18 ./configure 2013-08-18 03:29:20 https://gist.github.com/Heartmender/338bff09bc87b6dd277b 2013-08-18 03:29:49 https://gist.github.com/Heartmender/25cb77c73c29b56761e5 2013-08-19 04:00:43 https://bugs.freedesktop.org/show_bug.cgi?id=63747 2013-08-19 04:00:48 i just stumbled on that gem 2013-08-19 09:16:52 files from v2.4.11-80-g0f417dd uploaded 2013-08-19 09:20:38 files from v2.4.11-80-g0f417dd uploaded 2013-08-19 11:29:29 what's the plan with xen 4.3? 2013-08-19 11:31:19 shafire: i got it built but i need to rebase the patches 2013-08-19 11:31:30 unless kaniini beats me to it ;) 2013-08-20 07:05:13 hi 2013-08-20 07:05:49 ncopa: Yesterday I tried a cross-compiler :-) 2013-08-20 07:05:50 http://alpinelinux.org/comment/89#comment-89 2013-08-20 07:06:00 Are these typical uclibc errors now? 2013-08-20 07:08:50 dunno 2013-08-20 07:09:29 either gcc-ada is using those symbols and expect them from the libc 2013-08-20 07:09:52 yeah 2013-08-20 07:10:06 look at /home/dennis/src/build-gcc/gcc/ada/rts/adaint.c line 3797 2013-08-20 07:10:34 might be there are some #if ... #else ... #endif around there that goes wrong 2013-08-20 07:11:05 google __sched_cpualloc, find out what it does 2013-08-20 07:11:15 and look at the ada code why it is needed 2013-08-20 07:11:37 check if __sched_cpualloc is a gnu extension (i bet it is) 2013-08-20 07:11:45 then find out if there are a compatible way to do it 2013-08-20 07:11:50 that will work on uclibc 2013-08-20 07:12:14 ha 2013-08-20 07:12:21 guess what shows up if i google 2013-08-20 07:12:24 http://comments.gmane.org/gmane.comp.lib.uclibc.general/22736 2013-08-20 07:12:57 there is a message from you :D 2013-08-20 07:13:27 did you crosscompile uclibc too? 2013-08-20 07:13:37 a complete toolchain? 2013-08-20 07:13:39 I used --with-sysroot 2013-08-20 07:13:46 I hope that was right... 2013-08-20 07:14:28 there are tons of different ways to do it... im not sure exactly what you are doing... 2013-08-20 07:14:31 but i trust you do ;) 2013-08-20 07:14:37 well... i *hope* you do :) 2013-08-20 07:14:58 it might be it will work if you add # include 2013-08-20 07:15:02 in proper file 2013-08-20 07:15:33 without recompiling gcc? 2013-08-20 07:16:38 well i think your gcc is broken 2013-08-20 07:17:21 this is probably broken: /home/dennis/opt/cross/lib/gcc/x86_64-alpine-linux-uclibc/4.8.1/adalib/libgnat.a 2013-08-20 07:17:38 so you need to fix it in the C sources and recompile it 2013-08-20 07:17:44 So, I add the # include in adaint.c and recompile gcc to fix libgnat.a? 2013-08-20 07:18:03 possibly 2013-08-20 07:18:55 moh 2013-08-20 07:18:57 no 2013-08-20 07:19:39 hm 2013-08-20 07:19:53 i guess something else is wrong 2013-08-20 07:20:09 __sched_cpualloc shoudl be a macro in uclibc 2013-08-20 07:21:19 shafire: can you check if there are any #include in adaint.c? 2013-08-20 07:21:25 at the top 2013-08-20 07:21:39 i think thats the issue 2013-08-20 07:21:54 it is missing #include 2013-08-20 07:22:08 There was an #include , I will check again 2013-08-20 07:25:25 In line 3785 is #include 2013-08-20 07:26:27 hum 2013-08-20 07:26:30 then i dont know... 2013-08-20 07:30:51 http://nopaste.dk/p58730 2013-08-20 07:32:39 no 2013-08-20 07:32:41 better this one 2013-08-20 07:32:42 http://nopaste.info/ffcb612692.html 2013-08-20 07:39:29 but there is no stdlib.h 2013-08-20 07:42:59 but sched.h includes stdlib.h, or? 2013-08-20 07:44:02 so that's not the error 2013-08-20 08:11:37 ncopa: maybe my cross compiler can not take sched.h because of #if defined __USE_GNU && defined __UCLIBC_LINUX_SPECIFIC__ 2013-08-20 08:17:32 i dunno 2013-08-20 08:17:36 could be 2013-08-20 08:17:41 I think, my toolchain is wrong :D 2013-08-20 08:18:05 Do you know, how I can build the alpine-linux-uclibc toolchain on debian? 2013-08-20 08:18:20 I compiled a hello.c with alpine-gcc on my debian 2013-08-20 08:18:43 http://pastie.org/8252654 2013-08-20 08:18:49 not good 2013-08-20 08:24:42 http://bugs.alpinelinux.org/issues/1907 2013-08-20 08:25:08 Is Dubiousjim here? 2013-08-20 08:27:27 I don't think so 2013-08-20 10:15:31 i ordered wandboard quad for compiling arm stuff 2013-08-20 10:19:45 nice 2013-08-20 10:20:01 fabled: eta? 2013-08-20 10:20:52 in processing. i'd assume it be here in 3-5 working days as all items were in stock. 2013-08-20 10:21:14 so we have a working aports in 6 days? 2013-08-20 10:21:16 :) 2013-08-20 10:31:27 lol :D 2013-08-20 10:53:33 fabled: how many wandboard quads? 2013-08-20 10:54:36 where did you buy itß 2013-08-20 10:54:37 ? 2013-08-20 11:03:31 shafire, just one to start with. from mouser.com 2013-08-20 11:03:48 they had several in stock 2013-08-20 11:05:21 I see, 27x in stock 2013-08-21 08:28:55 clandmeter, delivery ETA for my arm builder is this Friday 6pm 2013-08-21 08:29:08 so maybe next week... 2013-08-21 08:29:26 fabled: nice :) but ill be on holidays... 2013-08-21 08:29:38 so im sure you'll have it working when im back 2013-08-21 08:29:43 perhaps 2013-08-21 08:29:57 if not, ill stay longer haha 2013-08-21 08:30:39 :-) 2013-08-21 08:30:48 where do you go for holidays? 2013-08-21 08:31:25 italy 2013-08-21 08:31:33 first stop venice 2013-08-21 08:31:55 ah, milano too? 2013-08-21 08:32:29 not sure yet. I think i prefer smaller places in italy 2013-08-21 08:33:03 If you like shopping >1000EUR/shirt, milano is the best place :D 2013-08-21 08:33:03 but its all open. we just see what we find. 2013-08-21 08:33:31 In milano, there is a cool hamburger restaurant called big mama or so 2013-08-21 08:33:41 If you like meat, you should give it a try :) 2013-08-21 08:33:49 im a tech guy. shopping is not my hobby. 2013-08-21 08:34:05 except online shopping of course :) 2013-08-21 08:34:35 but only on websites with regex matching against (pc|tech)* 2013-08-21 08:47:25 lol 2013-08-21 17:34:46 hi 2013-08-21 18:41:35 in which package can I find sched.h? 2013-08-21 18:42:38 I mean sched.c 2013-08-21 18:45:11 ncopa: Can you update to syslinux 6? 2013-08-22 07:35:51 ncopa: You read the message about syslinux? 2013-08-22 07:37:35 !last 2013-08-22 07:37:52 shafire: I guess he didn't 2013-08-22 07:47:28 shafire: hi 2013-08-22 07:47:30 sorry was out 2013-08-22 07:47:34 re syslinux 6? 2013-08-22 07:47:46 funny but i needed efi the other day 2013-08-22 07:47:47 yeah 2013-08-22 08:02:22 see you all in FOSDEM then ;-) 2013-08-22 08:04:51 bah, not sure if I'll be there in 2014 2013-08-22 08:05:45 maybe royger will, we've met this year 2013-08-22 08:21:51 Will you update to syslinux 6? 2013-08-22 08:22:07 maybe then will work syslinux over AMT KVM without the stupid mode:) 2013-08-22 08:31:07 i tried syslinux-6 but it was not that easy 2013-08-22 08:31:11 needs some work :-/ 2013-08-22 08:32:14 it fails to build 2013-08-22 08:33:32 mh ok 2013-08-22 08:33:53 ncopa: I got a working cross-compiler on debian to alpine (gnatmake works :)) 2013-08-22 08:34:04 really! 2013-08-22 08:34:06 cool! 2013-08-22 08:34:18 but that's stage 1 2013-08-22 08:34:26 stage 2 fails :( 2013-08-22 08:34:27 jeah 2013-08-22 08:34:38 stage1 is still progress :) 2013-08-22 08:34:53 so you can build helloworld.ada for alpine? 2013-08-22 08:34:56 yes 2013-08-22 08:35:19 very nice! 2013-08-22 08:35:44 then you need to cross-build gnatmake and those tools for alpine 2013-08-22 08:37:09 now with flag --build=$BUILD --host=$TARGET --target=$TARGET 2013-08-22 08:37:10 or? 2013-08-22 08:37:32 It fails constantly to build libiberty 2013-08-22 08:38:18 I was actually thinking about bootstrapping gnat from Arch, but I gave up when I realised that there is no clever way to put this in repo 2013-08-22 08:41:25 https://projects.archlinux.org/svntogit/packages.git/tree/trunk/PKGBUILD?h=packages/gcc 2013-08-22 08:41:27 I see 2013-08-22 08:41:34 --disable-install-libiberty 2013-08-22 08:41:40 Will that check out later :) 2013-08-22 08:44:38 barthalion: the idea is to build a bootstrap binary 2013-08-22 08:44:49 that we "publish", upload some place 2013-08-22 08:45:32 then we use the bootstrap gnat in APKBUILD 2013-08-22 08:45:40 ah 2013-08-22 08:46:14 Do you want to help me? :) 2013-08-22 08:47:11 next week, I'm using too old PC now to do any serious thing which require compiling 2013-08-22 08:47:23 still waiting for new laptop 2013-08-22 08:50:50 Ok. 2013-08-22 08:53:05 barthalion: Do you use Ada? 2013-08-22 08:56:00 I had no sleep the last days :D 2013-08-22 08:56:21 I do not, one user wanted to compile Ada program months ago 2013-08-22 08:56:39 I worked on the ada compiler from 19:00-3:00 the last 5 days 2013-08-22 08:57:36 I guess you are not married 2013-08-22 08:57:41 :D 2013-08-22 08:58:07 i would def get slapped for that. 2013-08-22 08:58:34 I am young :p 2013-08-22 09:00:04 yeah, compile while you are young and can ;) 2013-08-22 09:00:38 i suppose i had my share of 19:00-03:00 hacking... 2013-08-22 09:00:52 The bad is, it's compiling about 50min per round 2013-08-22 09:01:01 yes... 2013-08-22 09:01:11 and the failure is most times in 49min... 2013-08-22 09:01:13 lol 2013-08-22 09:01:22 it helps with fast computer, much ram and ssd disk 2013-08-22 09:01:30 i still dont have ssd here though 2013-08-22 09:01:32 then it takes only 40 min 2013-08-22 09:01:34 :p 2013-08-22 09:01:44 funny thing is 2013-08-22 09:01:57 often the autotools stupidness is slower than the actual compiling 2013-08-22 09:02:23 autotools can make even a quad core i7 slow... 2013-08-22 09:03:44 we just got new 128G msata ssd's here. there are 50% smaller then normal form-factor. really small... 2013-08-22 09:03:54 1.8"? 2013-08-22 09:03:59 cool 2013-08-22 09:04:15 maybe test out that flashcache thingy? 2013-08-22 09:04:26 actually i do have ssd on my private laptop 2013-08-22 09:04:33 Oh I need to test it too 2013-08-22 09:04:36 but the cpu there is not capable of any compiling 2013-08-22 09:04:40 you can get one for free if you want, if you buy brix i7 in benelux :) 2013-08-22 09:04:40 c7 2013-08-22 09:04:44 but I fight with the ada compiler at moment 2013-08-22 09:07:03 wow, there are 5 different projects on github called 'pingu' 2013-08-22 09:11:15 http://www.phison.com/Upload/PicFiles/2011.11.7_11.16.52_6142.jpg 2013-08-22 09:20:36 wow 2013-08-22 09:21:43 http://geizhals.at/de/mushkin-enhanced-atlas-deluxe-240gb-mknssdat240gb-dx-a755358.html 2013-08-22 09:22:04 that's my ssd in my server 2013-08-22 09:22:32 the flashcache device 2013-08-22 09:25:02 nice 2013-08-22 09:25:11 the oens we have are 50% smaller 2013-08-22 09:25:23 and they are cheap 2013-08-22 09:25:28 so i dont know the performance 2013-08-22 09:36:19 You got them for free? 2013-08-22 09:41:26 ncopa: funny thing is: the musl/arm ada gcc should be much simpler 2013-08-22 09:43:08 heh 2013-08-22 09:43:35 i wonder if it would be simpler to build a musl/x86 gcc statically 2013-08-22 09:44:39 I hope it will be 2013-08-22 09:51:07 If the ada compiler works, I hope, I get from you http://www.youtube.com/watch?v=S57LM0ULh-w :D 2013-08-22 10:11:22 mh 2013-08-22 10:11:33 ada programs which uses threads get killed... :( 2013-08-22 11:17:55 I should get a wandboard too 2013-08-22 11:18:14 fabled: could you tell me tomorrow, how the wandboard is? 2013-08-22 11:18:19 shafire, got it now 2013-08-22 11:18:42 but alas, wrong kind of power plug. it said 'UNIVERSAL' but that apparently refers to voltage input only - not the mechanics. 2013-08-22 11:19:28 wrong plug? 2013-08-22 11:19:37 yeah 2013-08-22 11:19:43 lol 2013-08-22 11:20:19 yes... we can build Alpine/ARM... oh wait, the power cord is wrong type. 2013-08-22 11:21:47 Could you get one in the electronic market 2013-08-22 11:21:49 or an adapter? 2013-08-22 11:27:16 yes 2013-08-22 11:27:30 need to get one later today. gotta finish some other things first. 2013-08-22 11:27:54 shafire, need to finish http://t2.fi/challenge/t213-challenge/ today 2013-08-22 11:28:34 :) 2013-08-22 11:29:27 how do you get the md5 hashes? 2013-08-22 11:29:43 do you need to go to the zetor restaurant? 2013-08-22 11:30:11 nah. as said: The challenge will be released on 2013-09-07 10:00 EEST. 2013-08-22 11:30:46 need to finish it... as in, i'm writing the challenge. 2013-08-22 11:30:50 ah 2013-08-22 11:30:51 lol 2013-08-22 11:30:52 sry 2013-08-22 11:30:59 Misread it 2013-08-22 11:34:03 using xen on wandboard, is this a good idea? 2013-08-22 11:37:50 ncopa: http://pastie.org/8259356 2013-08-22 11:37:53 do you have an idea 2013-08-22 11:37:57 this is stage 2 =) 2013-08-22 11:38:59 sounds like its looking for glibc 2013-08-22 11:39:11 or using glibc headers 2013-08-22 11:39:34 So, it will always look for it? 2013-08-22 11:40:53 Or I forgot the g++ compiler? 2013-08-22 11:40:59 i dunno 2013-08-22 11:41:18 how can gcc compile the g++ compiler? 2013-08-22 11:41:23 which language is this? c++? 2013-08-22 11:42:25 I will rebuild it with c++ activated... 2013-08-22 11:42:38 and check again 2013-08-22 11:43:59 I think, it's taking the wrong g++ 2013-08-22 12:58:21 no Mr Elegance for fabled this year :( 2013-08-22 13:27:34 If you have such wand board quad cards 2013-08-22 13:27:35 http://www.weltdergadgets.de/wp-content/uploads/2013/05/wandboard-quad.jpg 2013-08-22 13:27:40 How can you use them? 2013-08-22 13:31:08 ah, edm carrier board 2013-08-22 13:32:45 great 2013-08-22 13:32:50 http://www.seco.com/en/item/secocq7-mitx2_0/ <- with wandboard edm card 2013-08-22 13:32:51 cool 2013-08-22 13:33:45 oh, not an edm carrier board? :( 2013-08-22 13:40:56 http://www.data-modul.com/tl_files/images/Flyer/DM_Flyer_Embedded_0212_low.pdf <- the first board? 2013-08-22 13:52:41 another question... I have x86_64-...-gnatmake and a symbol link from gnatmake 2013-08-22 13:53:23 dennis@debian:~/HelloAda$ gnatmake hello.adb gcc -c hello.adb gcc: error trying to exec 'gnat1': execvp: No such file or directory gnatmake: "hello.adb" compilation error 2013-08-22 13:53:37 dennis@debian:~/HelloAda$ x86_64-alpine-linux-uclibc-gnatmake hello.adb x86_64-alpine-linux-uclibc-gcc -c hello.adb x86_64-alpine-linux-uclibc-gnatbind -x hello.ali x86_64-alpine-linux-uclibc-gnatlink hello.ali 2013-08-22 13:54:42 lrwxrwxrwx 1 dennis dennis 35 Aug 22 15:54 gnatmake -> x86_64-alpine-linux-uclibc-gnatmake 2013-08-22 13:54:53 what's the problem here? 2013-08-22 14:11:06 compiling solved the other g++ problem, I think 2013-08-22 14:11:13 compiling c++ solved the other g++ problem, I think 2013-08-22 15:47:43 jbilyk_: around? 2013-08-22 17:06:33 clandmeter: back 2013-08-23 09:26:02 files from v2.5.4-180-ge0bf949 uploaded 2013-08-23 21:12:32 hi 2013-08-23 21:12:51 does someone have some webspace, so I can upload the ada bootstrap? 2013-08-23 21:21:15 How much do you need? 2013-08-23 21:21:52 max. 500mb 2013-08-23 21:22:08 Just apk files and stuff? 2013-08-23 21:22:52 shafire: there's dev.alpinelinux.org/archive.alpinelinux.org 2013-08-23 21:22:59 shafire: apply for an account 2013-08-23 21:23:02 :D 2013-08-23 21:23:46 http://dev.alpinelinux.org/archive/ ? 2013-08-23 21:23:52 How can I apply? 2013-08-23 21:24:06 thanks for your help, Heartmender 2013-08-23 21:24:14 shafire: No problem ^_^ 2013-08-23 21:24:50 I should really look into making a rust package for Alpine 2013-08-23 21:25:26 I got Ada working :) 2013-08-23 21:26:18 shafire: write alpine-devel 2013-08-23 21:26:29 o 2013-08-23 21:26:31 ok 2013-08-23 22:53:34 Heartmender: Rust is self-hosted, or? 2013-08-23 22:53:45 shafire: It is 2013-08-23 22:53:59 shafire: Their reference compiler is linked to glibc :/ 2013-08-23 22:54:07 Could you not take the version, which was not self-hosted, compile it 2013-08-23 22:54:13 and compile the new version with it? 2013-08-23 22:54:24 I'd have to do that many times :/ 2013-08-23 22:54:54 why? 2013-08-23 22:55:14 Was 0.5 self hosted? 2013-08-23 22:55:16 The version before it was self-hosted can't compile a current rust compiler 2013-08-23 22:56:15 Ah, I see 2013-08-23 22:56:16 ok 2013-08-24 08:27:35 hi 2013-08-24 08:44:16 shafire: I don't have access to my ssh keys until monday or tuesday, so I can't push newer ferm 2013-08-24 08:44:30 ah ok 2013-08-24 08:44:32 np 2013-08-24 09:17:00 kaniini: I sent an email to devel 2013-08-24 09:17:15 ncopa will probably set you up 2013-08-24 09:17:37 ok 2013-08-24 18:16:42 Hi, there! I stumbled over a problem regarding the vim package. While busybox's vi handels utf-8 characters just fine, vim works in ascii mode by default. Adding "set encoding=utf-8" to "/etc/vim/vimrc" partially solves this problem. With this line added displaying files encoded in utf-8 works fine. But I still cannot input non-latin letters with my keyboard, which works perfectly in busybox's vi 2013-08-24 18:17:46 Personally I think as utf-8 is the default in /etc/profile it would be a good idea to add "set encoding=utf-8" to /etc/vim/vimrc by default 2013-08-24 18:18:04 Does anyone has an idea how I could solve the input problem? 2013-08-24 20:10:24 maribu: about the display issue, I haven't encountered such problems in edge nor in >= 2.5 2013-08-24 20:11:31 maribu: however, please send me a memo about adding it to default vimrc, on monday or tuesday 2013-08-24 20:12:12 maribu: not sure how to solve the input bug, but again, everything works fine here 2013-08-24 20:12:35 maribu: but I use Alpine only on remote machines, maybe it's a clue 2013-08-24 22:14:37 barthalion: Thank you for your answer! I'm using Alpine 2.6.3 with up to date vim package. It is said that vim defaults to using utf-8 without setting encoding to utf-8 if it detects that the locale is using utf-8 (e. g. env variable LANG=en_US.utf-8 is set). There might be a reason vim correctly detects that your locale is using utf-8, while it does not on my machine. 2013-08-24 22:15:03 I'll go to bed for now. Thanks for your time! 2013-08-25 07:07:15 clandmeter, kaniini, ncopa, rnalrd: rackspace is giving free cloud accounts to open source projects, are we interested? 2013-08-25 07:14:59 uhh 2013-08-25 07:15:04 considering they don't support 2013-08-25 07:15:10 alpine as an install image... 2013-08-25 07:15:12 lol 2013-08-25 07:16:23 what's the problem? we can make one 2013-08-25 07:19:59 well i idly wonder what we would do with it 2013-08-25 07:20:43 yeah, that's why I asked 2013-08-25 07:20:58 generally speaking we haven't any shortage of infrastructure 2013-08-25 13:03:10 hi 2013-08-25 15:23:31 please git pull git://git.alpinelinux.org/fab/aports 2013-08-25 15:23:54 there are upgrades for ansible and colletd 2013-08-25 15:23:56 thanks 2013-08-26 06:15:27 morning 2013-08-26 06:15:42 fabian_a: thanks! 2013-08-26 06:16:37 barthalion: re rackspace: good to know. I don't think we have any need atm. 2013-08-26 06:18:55 barthalion: do you have a link? 2013-08-26 06:20:34 ncopa: no, I can ask if they have more info somewhere 2013-08-26 10:10:57 ncopa: the binaries are ready 2013-08-26 10:11:11 ada-bootstrap 2013-08-26 10:11:22 I tried some complex ada programs and it's working :-) 2013-08-26 10:19:09 I am just waiting for the dev account to upload it :) 2013-08-26 11:48:43 is there any pkg for command line notes manager in al? 2013-08-26 11:51:42 vkrishn: command line notes manager? 2013-08-26 11:55:12 eg http://swapoff.org/DevTodo 2013-08-26 11:55:36 I think I would add to bugs.a.o with more links 2013-08-26 11:56:47 http://bugs.alpinelinux.org/issues/2227 2013-08-26 11:57:41 I think there needs to be one, but more options are always welcome 2013-08-26 11:59:39 look suseful 2013-08-26 11:59:59 :) 2013-08-26 12:04:46 I think there are most tools now in AL main+testing to build a personal centralized notes manager server, something that I was thinking for sometime :) 2013-08-26 12:06:34 combining search tools + light servers + php + ...etc 2013-08-26 14:00:05 hi 2013-08-26 14:01:50 the to do tool looks nice 2013-08-26 14:18:45 ncopa: how can I send you the ada bootstap binaries= 2013-08-26 14:18:46 ? 2013-08-26 20:26:58 ncopa: how do I release a new version of an ACF package 2013-08-26 20:27:11 seems you've changed the download URL because of problems with cgit 2013-08-26 20:54:06 ACTION is trying "abump acf-provisioning-0.3.3" and it can't find the package 2013-08-26 20:55:00 ACTION notices that acf-provisioning-0.3.3.tar.xz was (somehow) created on dev.a.o, but the APKBUILD file is looking for acf-provisioning-0.3.3.tar.bz2 2013-08-26 21:14:45 ncopa_: I changed the APKBUILD source to point to the .xz file 2013-08-26 21:14:52 please let me know if this is right / wrong way 2013-08-26 21:50:12 files from v2.5.4-183-g72257a0 uploaded 2013-08-26 21:50:18 files from v2.5.4-183-g72257a0 uploaded 2013-08-26 22:28:35 files from v2.4.11-82-g09a0ccf uploaded 2013-08-26 22:29:08 files from v2.4.11-82-g09a0ccf uploaded 2013-08-26 22:33:28 files from v2.3.6-261-g75137e0 uploaded 2013-08-27 05:53:25 tdtrask: looks correct 2013-08-27 07:45:41 hi 2013-08-27 08:29:08 i wonder if it would help to add more swapspace 2013-08-27 08:29:34 it appears to run out of memory when building openjdk6 2013-08-27 08:29:58 when both build-edge and build64-edge builds openjdk, one of them runs out of memory 2013-08-27 08:38:28 why not more ram? 2013-08-27 08:38:45 I don't use swap at all my installations 2013-08-27 08:39:24 how much ram do you have available? 2013-08-27 08:40:38 it already has 8G ram 2013-08-27 08:43:13 ncopa, there was a new version of acf-provisioning yesterday 2013-08-27 08:43:34 I can see the package if browsing to dl-3.a.o using a web browser 2013-08-27 08:43:53 but when using dl-3 in /etc/apk/repositories then the new package doesn't show 2013-08-27 08:44:06 Any clue on what might be wrong? 2013-08-27 08:44:09 apk update? 2013-08-27 08:44:13 Done 2013-08-27 08:44:15 apk version -U 2013-08-27 08:44:54 Interesting... It complains about 'Read-only file system' when running 'apk version -U' 2013-08-27 08:45:03 didn't complain when doing 'apk update' 2013-08-27 08:45:23 do you boot from cdeom? 2013-08-27 08:45:26 cdrom 2013-08-27 08:45:27 But when I change repository to e.g. nl.a.o, then I get the new package 2013-08-27 08:46:18 It's a CF card installation 2013-08-27 08:47:19 dl-3 and nl are inconsistent? 2013-08-27 08:47:32 Not that I can see 2013-08-27 08:47:37 just by browsing to them 2013-08-27 08:48:11 When changing to nl repository and running 'apk version -U' I still get the 'read-only file system' warning, but now I see the updated package 2013-08-27 08:48:38 Looks to me like a local version of the APKINDEX.tar.gz is broken (or something like it) 2013-08-27 08:48:39 they are consistent 2013-08-27 08:48:51 wget -q -O - http://nl.alpinelinux.org/alpine/edge/main/x86/APKINDEX.tar.gz | md5sum 2013-08-27 08:49:30 mhaveIa: I remember you having issues due to missing dirs in /var 2013-08-27 08:49:40 try apk fix alpine-baselayout 2013-08-27 08:50:18 I am using the 2.4 repository (not 'edge') 2013-08-27 08:50:38 When running the md5sum check on those repos, then nl and dl-3 are incosistent!!! 2013-08-27 08:52:01 could you please check what apk version --index say? 2013-08-27 08:52:21 Seems... it might be a proxy that holds a old version of the APKINDEX.tar.gz for one of the repos... 2013-08-27 08:52:35 but when I change repo, then proxy has nothing cached, and then things starts working 2013-08-27 08:52:51 could be 2013-08-27 08:52:52 so... seems it's due to proxy caching things 2013-08-27 08:53:28 Thanks for your help! 2013-08-27 08:54:19 http://www.cyberciti.biz/faq/lighttpd-send-cache-control-maxage-headers-of-staticfiles/ 2013-08-27 08:54:31 we should probably set cache control 2013-08-27 08:55:55 I guess, setting that on the mirrors would solve it... 2013-08-27 08:56:02 but can we force all mirrors to do that? 2013-08-27 08:56:17 Is there a way to get 'apk' to request a 'non-cached' version of the files? 2013-08-27 09:00:41 export http_proxy= ; apk ... 2013-08-27 11:37:28 ncopa, re: arm 2013-08-27 11:37:43 seems we need per-board version of u-boot binary as boot loader 2013-08-27 11:38:51 if upstream multiplatform kernel supports to board, then we can use single kernel+device tree blobs/per-board 2013-08-27 11:39:30 basic stuff of wandboard seems to be supported on linux-3.10.x, so i'm really happy 2013-08-27 11:39:46 rpi needs patches (perhaps even it's own kernel version) and the gpu binary blob 2013-08-27 11:40:37 throw away the old rpi :) 2013-08-27 11:42:21 hmm ok... 2013-08-27 11:42:42 shafire, it's quite new rpi. though, yes, it's old tech spec wise. 2013-08-27 11:43:23 i've been trying to boot grsec kernel on arm too, but that seems buggy still. 2013-08-27 11:43:33 http://forums.grsecurity.net/viewtopic.php?f=1&t=3720&p=13448 2013-08-27 11:44:03 coffee time. back after a while. 2013-08-27 11:51:43 algitbot: build master 2013-08-27 13:45:48 hm.. i wonder why build64-edge fails 2013-08-27 13:47:13 algitbot: build master 2013-08-27 13:55:15 :-( 2013-08-27 13:56:32 ncopa: thanks, I'll update the rest of the ACF packages as I need to upgrade them 2013-08-27 14:05:27 weird... openjdk appears to build on my desktop x86_64 2013-08-27 14:05:36 i dunno why it fails on build64-edge 2013-08-27 14:05:47 fabled: any ideas? 2013-08-27 14:05:58 out of ram? 2013-08-27 14:06:04 going for the day 2013-08-27 14:06:05 see you tm 2013-08-27 14:06:08 ok 2013-08-27 14:06:09 see u 2013-08-27 14:06:15 could be also parallel make issue 2013-08-27 14:06:42 ok i'll try dig into it 2013-08-27 14:13:24 i stopped qemu build-arm 2013-08-27 14:13:38 lighttpd on git.a.o was eating alot too 2013-08-27 14:13:49 algitbot: build master 2013-08-27 14:38:53 http://git.alpinelinux.org/cgit/aports/commit/?h=2.6-stable <- why is this commit in the future? 2013-08-27 14:39:58 shafire: someone's computer has the wrong datestamp and they committed 2013-08-27 14:40:06 shafire: likely tdtrask's ntp is busted 2013-08-27 14:41:53 ah okay 2013-08-27 14:54:49 jbilyk_: sorry, I did update it yesterday, but only after I saw the bad timestamps 2013-08-27 20:37:43 hi 2013-08-28 05:42:36 hm 2013-08-28 05:42:40 when is alpine 1.7 due? 2013-08-28 05:42:45 er 2013-08-28 05:42:46 2.7 2013-08-28 05:45:44 Mp5shooter, http://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases 2013-08-28 05:46:02 early November 2013-08-28 05:47:00 cool 2013-08-28 05:48:33 ncopa, ^^^shorewall-core fails 2013-08-28 05:52:13 anyone here have a juniper SRX? 2013-08-28 05:57:28 oh 2013-08-28 06:19:21 morning 2013-08-28 06:21:16 mornin 2013-08-28 06:21:26 goo dnight 2013-08-28 06:21:27 ;p 2013-08-28 06:57:51 gonna do xen 4.3 2013-08-28 06:57:53 today 2013-08-28 07:00:21 kaniini: nice! 2013-08-28 07:00:29 you can start with this: http://sprunge.us/NAZe 2013-08-28 07:00:30 or, at least try. 2013-08-28 07:00:45 it builds, but the websockets patches needs to be rebased 2013-08-28 07:01:10 yeah 2013-08-28 07:01:29 i need to upstream thoses 2013-08-28 07:01:35 -s 2013-08-28 07:01:52 as you see, there are some xsa's that still applies 2013-08-28 07:02:01 i dont know if they should be applied or not 2013-08-28 07:02:13 yeah 2013-08-28 07:02:33 maybe we could have royger give us a list of xsa patches that should be applied to 4.3 2013-08-28 07:02:36 yes 2013-08-28 07:02:44 well 2013-08-28 07:02:55 i have machines i can test on 2013-08-28 07:02:57 too 2013-08-28 07:53:30 hi 2013-08-28 07:56:17 any roadmap for gcc-gnat?:) 2013-08-28 07:56:38 will try look at it this week 2013-08-28 08:02:48 thanks 2013-08-28 09:19:06 Could you solve the jvm error from yesterday? 2013-08-28 12:13:16 kaniini: if you're about to do xen, could you please have a look at #2031 while there? 2013-08-28 12:13:40 you can grep checkpath */*.initd to find how other init.d scripts does it 2013-08-28 12:14:05 preferrible in separate commit so i can cherry-pick it for 2.6-stable 2013-08-28 12:54:17 you took task? :) 2013-08-28 13:12:34 shafire: lowest hanging fruit and it was not perl 2013-08-28 13:13:25 fabled: what do you think about: http://sprunge.us/OCNg 2013-08-28 13:13:44 apk search --has-origin APORTNAME... 2013-08-28 13:14:03 What's a lowest hanging fruit? 2013-08-28 13:14:34 "lowest hanging fruit" = "the easiest to pick" 2013-08-28 13:14:43 ncopa, looks ok to me 2013-08-28 13:15:00 maybe proofread the description 2013-08-28 13:15:29 and "from-origin" 2013-08-28 13:15:39 or just "origin" 2013-08-28 13:15:40 i think its a bit ugly to have them both in same func, whil beeing exlusive 2013-08-28 13:16:00 i mean. code is ok. just check the strings that are for help 2013-08-28 13:16:07 and the option name literal 2013-08-28 13:16:49 ncopa: thanks:) 2013-08-28 13:27:31 anyone native english speaking that can proofread the --has-origin line: http://sprunge.us/LEfV 2013-08-28 13:28:06 --has-origin List packages that has the given origin 2013-08-28 13:28:07 ? 2013-08-28 13:28:17 yes 2013-08-28 13:28:40 it prints the "origin" or the aport name 2013-08-28 13:29:26 example: https://dpaste.de/guFS0/ 2013-08-28 13:31:29 "maybe: List all the packages which have the given origin"? :) But I am not native... 2013-08-28 13:32:15 Lists* 2013-08-28 13:33:52 or "Print the packages which have the given origin" 2013-08-28 13:33:55 i dunno either :) 2013-08-28 13:35:16 I should really create my translate website (planning since 2007 - hehe)... 2013-08-28 13:36:00 speaking of... 2013-08-28 13:36:20 i would be interested in a database with all the .po translations 2013-08-28 13:36:50 .po translation of projects? 2013-08-28 13:36:52 so you coudl search how a string was translated to various applications 2013-08-28 13:36:59 good idea 2013-08-28 13:37:03 thanks! 2013-08-28 13:42:26 I've seen something like that 2013-08-28 13:42:35 but iirc it's only polish - english 2013-08-28 13:42:52 (and vice versa) 2013-08-28 13:43:54 transifex.com appears to do something like that 2013-08-28 13:46:21 my new laptop arrived, so I'll finally do more packaging 2013-08-28 13:46:29 congrats! 2013-08-28 13:47:47 barthalion: which laptop? 2013-08-28 13:48:07 nothing special, Lenovo w/ Celeron 1000M 2013-08-28 13:48:13 but still better than Pentium D 2013-08-28 13:48:14 :p 2013-08-28 13:48:45 enough to compile in chroot and test packages in virtual machine before push 2013-08-28 13:49:00 barthalion: i'd suggest using lxc 2013-08-28 13:49:24 is there tutorial to do in lxc? 2013-08-28 13:49:34 apk add lxc 2013-08-28 13:49:36 I want to make packages too: nodejs and mono 2013-08-28 13:49:41 ncopa: my main os is Arch Linux, so I usually use systemd containers 2013-08-28 13:49:51 shafire: nodejs is ready in testing 2013-08-28 13:50:01 just need pkgver bump and more testing 2013-08-28 13:50:12 ah, cool. okay 2013-08-28 13:50:20 lxc-create -t alpine -n mycontainer 2013-08-28 13:52:45 how to find out the exact list of aports to rebuild if mesa gets bumped and abi breaks: src/apk search --has-origin -q mesa | xargs apk search -r --exact -q --origin 2013-08-28 13:53:45 cool 2013-08-28 13:54:22 Hi. How/where do I request a new developer account on the GIT server? I have "lshw" and "scappy" packages I'd like to publish... 2013-08-28 13:54:38 how many git commits? 2013-08-28 13:55:05 if its 2 git commits, then git send-email --to alpine-devel@lists.alpinelinux.org 2013-08-28 13:55:19 so we can review 2013-08-28 13:55:49 if you have more than 5 commits then you can use github also 2013-08-28 13:56:44 paul_mezo: and if we like what you do and you stick around for a while, then we can create a dev account on git.a.o 2013-08-28 13:58:42 Hi. How/where do I request a new developer account on the GIT server? I have "lshw" and "scappy" packages I'd like to publish... 2013-08-28 13:58:56 15:54 <@ncopa> how many git commits? 2013-08-28 13:58:56 15:55 <@ncopa> if its 2 git commits, then git send-email --to alpine-devel@lists.alpinelinux.org 2013-08-28 13:58:59 15:55 <@ncopa> so we can review 2013-08-28 13:59:01 15:56 <@ncopa> if you have more than 5 commits then you can use github also 2013-08-28 13:59:04 15:57 <@ncopa> paul_mezo: and if we like what you do and you stick around for a while, then we can create a dev account on git.a.o 2013-08-28 13:59:43 cool, thanks! 2013-08-28 14:00:03 for now I have just two packages, but might have some more in the near future 2013-08-28 14:00:14 pkilar: very nice! 2013-08-28 14:00:26 pkilar: great 2013-08-28 14:00:27 pkilar: git send-email them for now 2013-08-28 14:00:39 will do, thanks! 2013-08-28 14:01:12 you can use github if you are ok with that too, and ask in here to pull 2013-08-28 14:01:35 pkilar: do you have experience with git? 2013-08-28 14:02:00 ncopa: thanks for 'task' pkg 2013-08-28 14:02:25 vkrishn: np, please test and feedback. i saw there was a bunch of scripts in the -doc pkg 2013-08-28 14:02:34 ok 2013-08-28 14:03:09 re: --has-origin 2013-08-28 14:03:10 List packages from the given origin 2013-08-28 14:03:15 perhaps 2013-08-28 14:03:20 ncopa: I'm more sysadmin guy that was "nicely asked" to do some development ;) 2013-08-28 14:03:31 or s/from/for/ 2013-08-28 14:03:46 pkilar: I guess you're not only case like that here 2013-08-28 14:03:49 my git knowledge is improving 2013-08-28 14:06:21 good 2013-08-28 14:20:07 ncopa: I got this: git: 'send-email' is not a git command. 2013-08-28 14:20:30 google says I need git-email package, but cannot find that package on Alpine 2013-08-28 14:20:46 git-perl iirc 2013-08-28 14:22:04 thanks! it worked 2013-08-28 14:23:11 is there something like send-email for hg? 2013-08-28 14:24:01 you don't want to use hg 2013-08-28 14:24:05 believe me 2013-08-28 14:24:50 :) 2013-08-28 14:25:13 btw, i was thinking of reorganizing git subpackages, and add a git-email package 2013-08-28 14:26:03 pkilar: http://wiki.alpinelinux.org/wiki/Creating_patches 2013-08-28 14:29:05 barthalion: why? 2013-08-28 14:31:02 it becomes freakin slow in big projects 2013-08-28 14:31:05 try to clone firefox 2013-08-28 14:31:49 one of Arch devs builds nightly versions, and cloning hg repo took a few days 2013-08-28 14:31:57 and would take more, but he switched to git mirror 2013-08-28 14:32:45 is it because of the family history? 2013-08-28 14:37:09 no idea 2013-08-28 14:46:17 sounds more like bad design 2013-08-28 14:46:28 or 'inferior' design 2013-08-28 14:50:08 hmpf we cannot upgrade ffmpeg to 2 til vlc supports it 2013-08-28 14:50:13 same with qt5 2013-08-28 14:50:38 well, we culd ship a qt5 package in addition to qt (4) 2013-08-28 14:51:37 rename qt to qt4 2013-08-28 14:51:54 and link qt to qt4? 2013-08-28 14:51:58 or to complicated? 2013-08-28 14:52:01 too* 2013-08-28 14:52:32 ncopa: you can't update to qt5 anyway 2013-08-28 14:52:47 ncopa: I'd rename qt to qt4 and add qt5 2013-08-28 14:52:53 ncopa: e-mail w/patches sent... 2013-08-28 14:53:16 ncopa: about ffmpeg... in Arch we use ffmpeg-compat based on older branch, only for some packages 2013-08-28 14:53:59 i saw 2013-08-28 14:54:31 i think we only have vlc (and possibly xbmc) 2013-08-28 14:54:34 atleast in main 2013-08-28 14:54:40 or you can update vlc to prerelease 2013-08-28 14:54:47 this is what Fedora devs did in rawhide 2013-08-28 14:55:00 seems like vlc-2.1 is around the corner 2013-08-28 14:56:10 i have no hurry 2013-08-28 14:57:37 pkilar: looks ok. i'd prefer separate commits for each package 2013-08-28 14:57:45 commit message should be: 2013-08-28 14:57:56 testing/: new aport 2013-08-28 14:58:07 then a link to url and a short desc 2013-08-28 14:58:18 ok 2013-08-28 14:58:22 will do 2013-08-28 14:58:35 do you want retry? or I commit as is? 2013-08-28 14:58:50 up to you 2013-08-28 15:00:56 oh.. 2013-08-28 15:00:57 no 2013-08-28 15:01:01 depends_dev 2013-08-28 15:01:04 is wrong 2013-08-28 15:01:12 ok i gotta go 2013-08-28 15:01:20 depends_dev= should be empty 2013-08-28 15:01:29 and makedepends="sqlite-dev" 2013-08-28 15:01:33 and thats it i think 2013-08-28 15:04:18 i gotta run 2013-08-28 15:04:26 see u 2013-08-28 15:07:56 see you 2013-08-28 15:27:24 ncopa: please disregard my e-mail, sent wrong patch :( 2013-08-28 20:09:31 hi 2013-08-28 20:09:42 the url of the package scampy is incorrect 2013-08-28 20:09:42 http://git.alpinelinux.org/cgit/aports/commit/?id=f2fbff6ef653ee27fc8451c438b3363a89ee7132 2013-08-28 20:10:36 sorry 2013-08-28 20:10:38 I mean http://git.alpinelinux.org/cgit/aports/commit/?id=5df815a4992c60cd0496e46c0b391cb2b5aca47a 2013-08-28 20:10:46 url should be https://www.secdev.org/scapy/ 2013-08-28 20:11:03 oups 2013-08-28 20:11:18 http://www.secdev.org/projects/scapy/ is working now, ... 2013-08-28 20:11:51 As you were! 2013-08-28 20:11:53 :) 2013-08-28 21:42:49 barthalion: I cloned firefox with git and hg 2013-08-28 21:43:06 hg (20min) was faster than git (25min) 2013-08-28 21:43:07 mh? 2013-08-29 06:03:46 hi 2013-08-29 06:03:58 fabled: i pushed to apk-tools yesterday 2013-08-29 06:04:16 ncopa, great 2013-08-29 06:11:01 fabled: maybe do rc3? 2013-08-29 06:11:23 ncopa, i have few commits pending too. but yes, should do rc3 soon 2013-08-29 06:46:37 would there be a need for adding a library for EDI ? 2013-08-29 06:46:43 EDI? 2013-08-29 06:46:47 or apps add their own libs 2013-08-29 06:46:49 https://github.com/nevali/libedi 2013-08-29 06:47:21 who/what uses libedi? 2013-08-29 06:48:17 yes, I think individual apps add them 2013-08-29 06:48:32 or inbuild it 2013-08-29 06:49:26 fabled: i am looking at the crosscompiled gcc-gnat that shafire built 2013-08-29 06:49:34 it ships a complete gcc 2013-08-29 06:49:59 now i'd like to only use gcc-gnat from there to build gcc 2013-08-29 06:50:02 not the entire gcc 2013-08-29 06:50:35 http://sprunge.us/VHaO 2013-08-29 06:52:05 and the build gnatmake is broken 2013-08-29 06:52:08 dev32-edge:~/aports/main/gcc/src/gcc-gnat-alpine-x86/bin$ ./gnatmake 2013-08-29 06:52:08 sh: ./gnatmake: not found 2013-08-29 06:53:11 i think the INTERP header is wrong 2013-08-29 06:53:15 https://dpaste.de/xnLy4/ 2013-08-29 07:10:16 ncopa, any plans to publish git repo stats ? 2013-08-29 07:10:18 http://scm-pepper.sourceforge.net/ 2013-08-29 07:10:40 http://git.alpinelinux.org/cgit/aports/stats/ 2013-08-29 07:10:48 :) 2013-08-29 07:11:43 no fancy graphs 2013-08-29 07:12:22 patches are welcome ;) 2013-08-29 07:12:46 fancy graphs is not a priority for me atm 2013-08-29 07:14:58 so would be ok to add pepper with generating scripts ? or cgit patch ? 2013-08-29 07:17:41 adding pepper is a good start 2013-08-29 07:18:28 speaking of fancy stats 2013-08-29 07:18:32 http://www.youtube.com/watch?v=61crugfsfDg 2013-08-29 07:18:48 ok, I would download and setup and see if its easy to add 2013-08-29 07:19:00 to add on git.a.o 2013-08-29 07:46:32 ncopa, busy? i have lua5.2 module request 2013-08-29 07:46:42 ok 2013-08-29 07:46:49 this one https://github.com/zhaozg/lua-openssl -- while there are releases, they are old, so it'd need to be git snapshot 2013-08-29 07:46:56 ok 2013-08-29 07:47:00 I'll look at it 2013-08-29 07:47:07 thanks! 2013-08-29 07:53:30 hi 2013-08-29 07:53:34 hi 2013-08-29 07:53:41 i'm compiling gcc with gnat now.. 2013-08-29 07:53:48 :)) 2013-08-29 07:53:52 your bootstrap gnat was not perfect 2013-08-29 07:54:03 :-( 2013-08-29 07:54:11 What was missing or wrong? 2013-08-29 07:54:11 but i think good enough for building a new bootstrap gnat 2013-08-29 07:54:58 it had interp header wrong 2013-08-29 07:55:07 https://dpaste.de/xnLy4/ 2013-08-29 07:55:21 [Requesting program interpreter: /opt/alpine-x86/lib/ld-uClibc.so.0.9.32] 2013-08-29 07:55:41 i worked around it my creating a symlink /opt/alpine-x86/lib that points to /lib 2013-08-29 07:56:11 then second issue was that it used libmpc.so.2 2013-08-29 07:56:35 did you: ln -s /opt/alpine-x86 / ? 2013-08-29 07:56:43 i rebuilt the old mpc on my system 2013-08-29 07:56:44 no 2013-08-29 07:56:58 ln -s /lib /opt/alpine-x86/lib 2013-08-29 07:57:01 wrong: ln -s / /opt/alpine-x86 2013-08-29 07:57:14 heh 2013-08-29 07:57:21 that was enough apparently 2013-08-29 07:57:39 it's because of --with-sysroot compiling 2013-08-29 07:57:45 i think it only needed the ld-uClibc.so.0.9.32 2013-08-29 07:57:46 ues 2013-08-29 07:57:49 yes i figured 2013-08-29 07:58:15 oh well... 2013-08-29 07:58:26 java fails now... 2013-08-29 07:59:01 no 2013-08-29 07:59:11 its not java 2013-08-29 07:59:13 its ada 2013-08-29 07:59:19 whats missing? 2013-08-29 08:00:00 https://dpaste.de/6G5Jg/ 2013-08-29 08:00:29 its my build ada thats broken 2013-08-29 08:00:34 due to hardened 2013-08-29 08:00:38 will have to check that up 2013-08-29 08:02:51 You should also check the libiberty lib or it will fail 5 minutes before finishing 2013-08-29 08:08:52 because of the error message, I found http://flatassembler.net/ :O 2013-08-29 08:27:34 flatassembler 2013-08-29 08:27:39 has gtk examples... 2013-08-29 08:30:16 It reminds me of Roller Coaster Tycoon 2, which was written in asm 2013-08-29 08:38:02 ncopa: Have you found a solution? 2013-08-29 08:38:10 no working on other thing now 2013-08-29 08:38:17 will look at how gentoo hardenend does it 2013-08-29 08:54:38 fabled: lua-openssl is a mess 2013-08-29 08:54:46 oh? 2013-08-29 08:54:49 mix crlf,lf line endings 2013-08-29 08:55:07 no support for lua 5.2 2013-08-29 08:55:11 i am porting it 2013-08-29 08:55:12 but 2013-08-29 08:55:15 it should have lua 5.2 2013-08-29 08:55:16 in git head 2013-08-29 08:55:24 this is git head... 2013-08-29 08:55:44 https://github.com/zhaozg/lua-openssl/commit/6620773540f8a9c6db5e1e092c4db52dcb0730a4 2013-08-29 08:55:54 maybe there's more stuff broke then 2013-08-29 08:56:00 hmmm 2013-08-29 08:56:32 or later commits broke 5.2 support again 2013-08-29 08:56:38 weird 2013-08-29 08:56:47 that does not look like mine... 2013-08-29 08:56:50 i'll dig into it 2013-08-29 08:57:08 i also got this: https://dpaste.de/fyF92/ 2013-08-29 08:57:41 underlinks? needs -lcrypto 2013-08-29 08:58:05 or just broken 2013-08-29 08:58:08 dev32-edge:~/src/lua-openssl$ readelf -d openssl.so | grep NEED 2013-08-29 08:58:08 0x00000001 (NEEDED) Shared library: [libcrypto.so.1.0.0] 2013-08-29 08:58:08 0x00000001 (NEEDED) Shared library: [libssl.so.1.0.0] 2013-08-29 08:58:11 just broken 2013-08-29 08:59:40 sounds like those are old openssl things, and removed as obsolete 2013-08-29 09:00:27 or... 2013-08-29 09:00:36 those are private openssl funcs that are not exported 2013-08-29 09:00:43 (but broken windows exports them) 2013-08-29 09:01:06 this breaks lua 5.2 again: https://github.com/zhaozg/lua-openssl/commit/e39b9afeb8f3a98e29644be7db47f16aa29722c6 2013-08-29 09:01:07 i think they recently fixed symbol visibility in openssl, so that might trigger it 2013-08-29 09:01:37 in general, it is a mess 2013-08-29 09:01:49 looks like so 2013-08-29 09:01:58 but it's the openssl bindings that are most complete 2013-08-29 09:02:30 hum 2013-08-29 09:02:32 https://github.com/mkottman/luacrypto might be okay too 2013-08-29 09:03:48 yup, that would be okay too 2013-08-29 09:03:53 ok goody 2013-08-29 09:03:59 i'll do that instead 2013-08-29 09:04:04 oh 2013-08-29 09:04:12 lua-crypto is in testing, but the original 2013-08-29 09:04:17 yeah 2013-08-29 09:04:18 the fork provides 0.3.0 2013-08-29 09:04:23 i'm gonna fix it 2013-08-29 09:04:25 ok. thanks. 2013-08-29 09:04:25 and do lua5.2 2013-08-29 09:04:32 i'm off to lunch 2013-08-29 10:52:59 shafire: Why on earth would you want hardened ada? 2013-08-29 10:53:13 i asked about ada on gentoo-hardened 2013-08-29 10:53:55 mh, I want to run ada on alpinelinux 2013-08-29 10:54:00 Or why he is asking? 2013-08-29 10:55:57 i think he wonder why you want run ada at all 2013-08-29 11:00:00 I want to compile and run ada programs on alpinelinux. I can't say more...:S 2013-08-29 11:01:27 I am out for lunch, back in 1h 2013-08-29 11:26:37 \o/ 2013-08-29 11:27:27 libdecnumber needs to be compiled with -fPIC 2013-08-29 11:27:53 bah 2013-08-29 11:27:54 no 2013-08-29 11:48:03 there is no package named libdecnumber? 2013-08-29 11:48:11 or even number 2013-08-29 11:53:20 there is a libdecnumber.a 2013-08-29 11:53:32 but it does not help compiling it with -fPIC 2013-08-29 11:53:46 the strange thing is that gnatmake gives error 2013-08-29 11:53:50 but not gnatlink 2013-08-29 11:53:59 and they are built very similar 2013-08-29 11:54:12 i wonder if i should give up and just add -fno-PIE 2013-08-29 11:54:58 mh 2013-08-29 11:56:03 where is libdecnumber.a? /lib or /usr/lib or /usr/include 2013-08-29 11:57:58 in gcc sources 2013-08-29 11:58:03 but that was not the problem 2013-08-29 11:58:12 aha 2013-08-29 11:58:18 same thing does happen with gnatlink 2013-08-29 12:10:08 yup 2013-08-29 12:10:14 adding -fno-PIE works 2013-08-29 12:11:38 but i iget textrels in the end... 2013-08-29 12:17:52 ACTION <- back @home, married ;-) 2013-08-29 12:18:06 hi all together! 2013-08-29 12:19:22 hi 2013-08-29 12:19:28 dev32-edge:~/aports/main/gcc/src/build$ scanelf -qT gcc/gnatkr 2013-08-29 12:19:28 gnatkr: execute_stack_op [0x2A8F] in (optimized out: previous _Z8ggc_freePv) [0x2A8E] 2013-08-29 12:19:28 gcc/gnatkr 2013-08-29 12:22:41 hm 2013-08-29 12:22:45 its libgcc 2013-08-29 12:26:29 StarWarsFan: :) how is your new girl? :) 2013-08-29 12:28:03 ncopa: libgcc_s.so.1? 2013-08-29 12:28:12 no 2013-08-29 12:28:24 it links to static libgcc apparently 2013-08-29 12:29:53 execute_stack_op 2013-08-29 12:30:09 it could be libbacktrace too 2013-08-29 13:05:54 ok i think it built now 2013-08-29 13:12:11 lost the connection 2013-08-29 13:12:43 09:05 ok i think it built now 2013-08-29 13:12:50 built? it's ready? 2013-08-29 13:13:09 i think it built and i think it will work 2013-08-29 13:13:24 :), what have you changed? 2013-08-29 13:13:32 what i need to do now, is build a new boot-strap gnat 2013-08-29 13:13:41 and then craetea bootstrap-gnat apkbuild 2013-08-29 13:13:50 which finally will be a dependency for gcc 2013-08-29 13:14:03 i added -fno-PIE too all the gnat* tools 2013-08-29 13:14:03 Ok 2013-08-29 13:14:34 What's PIE? I did not have this error, when I built it 2013-08-29 13:14:37 just PIC 2013-08-29 13:14:59 position independent executable 2013-08-29 13:15:10 needed to make full use of the PaXe'd kernel 2013-08-29 13:15:42 i am not yet sure, but it might be that -fno-PIE is needed on all ada programs too 2013-08-29 13:15:53 and if that is the case, then we will have to add it to the spec somehow 2013-08-29 13:16:00 to the gcc specs 2013-08-29 13:16:20 http://docs.adacore.com/examples/gnat-examples.tgz 2013-08-29 13:16:33 gnatmake -P.gpr 2013-08-29 13:16:35 To test it 2013-08-29 13:16:58 i need to rebuild gcc first 2013-08-29 13:17:00 with the patch 2013-08-29 13:17:19 and that splits the libgnat and gcc-gnat subpackages 2013-08-29 13:18:01 Ok, I am happy :) 2013-08-29 13:22:06 are ada runtimes linked to libgnat-4.8.so? 2013-08-29 13:22:09 and libgnarl? 2013-08-29 13:23:09 I don't think so 2013-08-29 13:23:17 You mean, the compiled programs? 2013-08-29 13:23:54 yes 2013-08-29 13:24:08 My Hello World was just linked against uclibc 2013-08-29 13:24:14 ok 2013-08-29 13:24:23 so i dont bother to create a libgnat package 2013-08-29 13:27:00 http://packages.debian.org/wheezy/libgnat-4.6 2013-08-29 13:27:05 Looks like it is sometimes needed? 2013-08-29 13:28:50 yeah, i'd just go ahead and split it 2013-08-29 15:18:17 http://stackoverflow.com/questions/2463150/fpie-position-independent-executable-option-gcc-ld 2013-08-29 15:18:21 Note, that PIE may be incompatible with -static 2013-08-29 15:22:59 ncopa: maybe the -static is a problem in https://dpaste.de/6G5Jg/ 2013-08-29 15:23:00 ? 2013-08-30 06:58:55 urgh 2013-08-30 06:59:04 i think we really dont want ada/gnat 2013-08-30 06:59:16 it will be a major pain to maintain it 2013-08-30 07:23:16 hi 2013-08-30 07:23:31 ncopa: okay 2013-08-30 07:23:45 I will use another distro for my ada stuff 2013-08-30 07:24:00 what i can do though 2013-08-30 07:24:01 is it because of uclibc or grsecurity? 2013-08-30 07:24:05 no 2013-08-30 07:24:06 not really 2013-08-30 07:24:13 that part works actually 2013-08-30 07:24:19 the problem is the bootstrap thingy 2013-08-30 07:24:32 i thought i could make a split pkg and use that as bootstrap gcc 2013-08-30 07:24:38 bootstrap ada 2013-08-30 07:24:45 together with current gcc 2013-08-30 07:24:49 but that wont work 2013-08-30 07:25:15 we'll need the entire bootstrap-ada-gcc to build our gcc - always 2013-08-30 07:25:42 unless we manually want bootstrap every new buildserver we set up (each new release every 6 months) 2013-08-30 07:25:55 however 2013-08-30 07:26:02 since we already made it this far 2013-08-30 07:26:16 i think i'll add halfway support to ada 2013-08-30 07:26:25 that is, there will be support for it in the APKBUILD 2013-08-30 07:26:51 but we will not ship/maintain a gnat from official builds 2013-08-30 07:27:15 but you should be able to build a working gcc-gnat yourself from sources 2013-08-30 07:27:29 with: LANG_ADA=true abuild -r 2013-08-30 07:27:36 okay 2013-08-30 07:27:47 ...given that you have a working bootstrap ada in PATH 2013-08-30 07:28:06 i'll ship the patch we need 2013-08-30 07:28:10 and i'll test it 2013-08-30 07:28:36 Is it, because it's self hosted? 2013-08-30 07:28:42 yes 2013-08-30 07:28:42 Will Rust have the same problem? 2013-08-30 07:29:05 can you build a bootstrap rust with other lang? 2013-08-30 07:29:12 we have the same issue with java too actually 2013-08-30 07:29:15 and it is a pain 2013-08-30 07:29:35 but with java, we can build a bootstrap java implemented in C (gcc-java) 2013-08-30 07:29:43 and use that to bootstrap openjdk 2013-08-30 07:30:04 with ada there are no bootstrap ada implemented in C 2013-08-30 07:30:12 that can be used 2013-08-30 07:30:14 hm, not sure about rust 2013-08-30 07:30:23 How about a non-alpine buildserver who actually built us the bootstraps? 2013-08-30 07:30:35 build* 2013-08-30 07:30:44 the problem is that we will depend on those 2013-08-30 07:30:48 for everything 2013-08-30 07:30:57 i mean for our basic, default gcc 2013-08-30 07:31:22 in other words, we would end up in a situation where we depend on N other build servers to be able to build gcc 2013-08-30 07:31:44 it just gets too complicated 2013-08-30 07:32:23 if there was a self-bootstrap ada script 2013-08-30 07:32:50 ncopa: do you mean something like this 2013-08-30 07:32:56 let me look for it 2013-08-30 07:33:01 i dont think there is 2013-08-30 07:33:06 I found something wait 2013-08-30 07:33:37 basically, the problem is that its too much work to maintain it 2013-08-30 07:34:54 https://dpaste.de/dmjX3/ 2013-08-30 07:35:20 but it has the same problem "you have GCC and the current or the previous version of GNAT installed" 2013-08-30 07:36:08 i think dubiousjim had similar issue with other language 2013-08-30 07:36:14 haskell or what it was 2013-08-30 07:36:46 hm 2013-08-30 07:37:11 ok 2013-08-30 07:37:15 i gotta think about it 2013-08-30 07:37:34 it is actually not worse than force ada in addition to c,c++ 2013-08-30 07:37:37 Actually, we can only have C and C++? Or? 2013-08-30 07:37:43 yes 2013-08-30 07:37:47 c,c++ 2013-08-30 07:37:51 c++ is required too 2013-08-30 07:37:51 And any self-hosted language is a pain? 2013-08-30 07:37:57 yes 2013-08-30 07:38:01 well 2013-08-30 07:38:14 only as long as it is difficult to bootstrap 2013-08-30 07:38:27 is there are simple ways to bootstrap it, then no problem 2013-08-30 07:38:29 like perl i thin 2013-08-30 07:38:36 you need perl to build perl 2013-08-30 07:38:44 but when building perl, you first build a miniperl 2013-08-30 07:38:54 which is then used to build the real perl 2013-08-30 07:40:29 but really 2013-08-30 07:40:40 those languages in gcc 2013-08-30 07:40:48 that are selfhosting 2013-08-30 07:40:55 like C and c++ 2013-08-30 07:40:57 and ada 2013-08-30 07:41:08 its just a question of adding them in first place 2013-08-30 07:41:17 when we build the first gcc for an arch 2013-08-30 07:41:39 if we do that, then there is no problem 2013-08-30 07:42:04 (well, only same problem as with c and c++ itself) 2013-08-30 07:42:09 ok 2013-08-30 07:42:13 here is what we do 2013-08-30 07:42:22 i'll add what i got so far to the APKBUILD 2013-08-30 07:43:07 i think it will let you build gcc-gnat.apk with a bootstrap ada 2013-08-30 07:43:14 then i'll think of it 2013-08-30 07:43:54 specially now that we are in the process of setting up arm/musl 2013-08-30 08:07:53 ncopa: a external buildserver? 2013-08-30 08:07:59 an external buildserver? 2013-08-30 08:09:01 no, i dont want depend on external build servers 2013-08-30 08:09:06 that adds complexity 2013-08-30 08:11:17 How about a glibc.apk for the build process? 2013-08-30 08:12:12 1) install glibc.apk 2) build ada-bootstrap 3) delete glibc.apk 4) build ada 2013-08-30 08:12:24 i dont think that will work 2013-08-30 08:12:37 douldnt the resulted ada depend on glibc? 2013-08-30 08:13:46 That's how I built the ada for alpine on Debian 2013-08-30 08:14:10 --host and --target to alpine (and the dynamic linker patch) 2013-08-30 08:15:00 thats how it will be needed to be done i think 2013-08-30 08:15:19 but i dont think we should separate the ada bootstrap from c/c++ bootstrap 2013-08-30 08:15:24 we should do them both at same time 2013-08-30 08:15:39 Ok 2013-08-30 08:15:46 What's about the other gcc languages? 2013-08-30 08:15:53 Are they not self-hosted? 2013-08-30 08:15:57 these* 2013-08-30 08:15:57 atleast if we want official support gcc-gnat 2013-08-30 08:16:10 you mean like java and fortran and go? 2013-08-30 08:16:12 no, they are not 2013-08-30 08:16:13 yes 2013-08-30 08:16:30 you can build those once you have a working c/c++ compiler 2013-08-30 08:17:37 mh, then lets put ada into the c/c++ bootstrap 2013-08-30 08:17:40 :) 2013-08-30 08:18:17 that woudl be the simplest 2013-08-30 08:19:12 Is then ada preinstalled? 2013-08-30 08:19:33 or will it be a gcc-gnat package? 2013-08-30 08:24:10 it will be a gcc-gnat package 2013-08-30 08:24:25 but hum 2013-08-30 08:24:35 it will introduce a circular dependency... 2013-08-30 08:24:47 gcc makedepends on gcc-gnat 2013-08-30 08:58:13 sweet, i think my APKBUILD works 2013-08-30 08:59:33 this feeling when package builds 2013-08-30 08:59:47 better, has been built 2013-08-30 08:59:56 priceless 2013-08-30 09:00:01 exactly 2013-08-30 09:00:24 the thing is, it actually works with shafires' bootstrap gnat 2013-08-30 09:00:31 Are you still using my gnat? 2013-08-30 09:00:37 oh 2013-08-30 09:00:40 maybe i wasnt 2013-08-30 09:00:45 yes i think i did 2013-08-30 09:01:02 no 2013-08-30 09:01:06 i think i was not 2013-08-30 09:01:10 but i think it will work 2013-08-30 09:01:32 use the gnat examples 2013-08-30 09:01:48 i have already built helloworld 2013-08-30 09:01:53 and gcc-gnat with it 2013-08-30 09:17:39 shafire: http://dev.alpinelinux.org/~ncopa/packages/x86/ 2013-08-30 09:18:23 :-) 2013-08-30 09:20:45 Will the packages stay there? 2013-08-30 09:20:54 no 2013-08-30 09:21:25 i think the plan is to build a similar pkg for x86_64 2013-08-30 09:21:37 yes 2013-08-30 09:21:42 then manually install gnat to build servers 2013-08-30 09:21:52 and finally update the apkbuild to enable it by default 2013-08-30 09:22:06 in other words bootstrap the build servers 2013-08-30 09:22:15 once they are bootstrapped we should be ok 2013-08-30 09:22:29 So you built ada without using my gnat compiler? 2013-08-30 09:22:32 How you did that? 2013-08-30 09:22:48 i built first ada with your gnat compiler 2013-08-30 09:22:57 to build a gnat.apk 2013-08-30 09:23:19 Is the -fno-PIE still there? 2013-08-30 09:23:20 then i used that to test the recent gcc/APKBUILD change 2013-08-30 09:23:23 yes 2013-08-30 09:23:25 i just pushed it 2013-08-30 09:23:50 http://git.alpinelinux.org/cgit/aports/commit/?id=81a0e21b890d05d0012d2c6a16e955bdca66da49 2013-08-30 09:24:39 now it should be possible to: export PATH=$path_to_shafires_ada/bin:$PATH ; LANG_ADA=true abuild -r 2013-08-30 09:24:53 i will try soon with x86_64 2013-08-30 09:25:00 i have some other work to doo meanwhile 2013-08-30 09:25:14 but i think we are seeing the light at the end of the tunnel 2013-08-30 09:25:42 i talked with fabled, and he appears to be ok with adding ada to the initial toolchain bootstrap 2013-08-30 09:25:54 (he is working on arm/musl) 2013-08-30 09:26:30 :-) 2013-08-30 09:28:18 i should fix the sec issues now 2013-08-30 09:32:47 shafire, ncopa, as long as it cross-compiles. i'm good :) 2013-08-30 09:37:47 fabled: May you contact me, when you are ready? 2013-08-30 09:58:21 shafire, yes? 2013-08-30 09:58:41 thanks 2013-08-30 09:58:42 oh you mean, when i've cross-compiled it? 2013-08-30 09:58:57 or? 2013-08-30 09:59:23 A working alpine on arn 2013-08-30 09:59:25 arm* 2013-08-30 10:00:12 oh 2013-08-30 10:00:25 it's work-in-progress 2013-08-30 10:00:32 i have working initramrd 2013-08-30 10:00:51 but i need to rebootstrap it due to some recent changes in various places 2013-08-30 11:30:01 files from v2.5.4-185-g4b3ee90 uploaded 2013-08-30 11:30:06 files from v2.5.4-185-g4b3ee90 uploaded 2013-08-30 11:33:39 files from v2.4.11-83-g2b610e7 uploaded 2013-08-30 11:34:00 files from v2.3.6-262-g921298d uploaded 2013-08-30 11:34:10 files from v2.3.6-262-g921298d uploaded 2013-08-30 12:02:11 files from v2.3.6-263-g6b1166a uploaded 2013-08-30 12:02:13 files from v2.3.6-263-g6b1166a uploaded 2013-08-30 12:02:25 files from v2.5.4-186-g54916c0 uploaded 2013-08-30 12:11:52 files from v2.4.11-87-g9341808 uploaded 2013-08-30 12:13:14 i think i'll do 2.6.4 on tuesday 2013-08-30 12:13:35 files from v2.4.11-88-gc88fc5c uploaded 2013-08-30 12:14:53 Will be there OpenJDK 1.7 in the next time? 2013-08-30 12:15:39 not for any 2.6.x 2013-08-30 12:18:13 files from v2.4.11-87-g9341808 uploaded 2013-08-30 12:24:52 heh, okay 2013-08-30 13:04:13 files from v2.5.4-187-g4a619cd uploaded 2013-08-30 13:04:22 files from v2.5.4-187-g4a619cd uploaded 2013-08-30 13:05:09 files from v2.3.6-264-g941eceb uploaded 2013-08-30 13:05:09 files from v2.4.11-89-gba80ac3 uploaded 2013-08-30 13:05:10 files from v2.4.11-89-gba80ac3 uploaded 2013-08-30 13:05:10 files from v2.3.6-264-g941eceb uploaded 2013-08-30 14:08:07 ncopa: ERROR: lua5.2-llthreads: all failed 2013-08-30 14:08:22 i'm on it 2013-08-30 14:08:26 tnx 2013-08-30 14:08:53 bah... 2013-08-30 14:08:59 it only finds lua5.1