**** BEGIN LOGGING AT Tue Apr 28 03:00:03 2009 Apr 28 03:15:30 Anyone have a minute for a newbie question? I'm attempting to install sweetlilmre's openzipit build as outlined here: http://sweetlilmre.blogspot.com/2009/04/new-oe-build.html. I've created a vfat partition for zImage-zipit2.bin and an ext3 partition for wireless-image-zipit2.tar. I've successfully untarred the tarfile, but can't seem to get zImage to take. I had the crazy idea that I could just dd the bin image to /dev/sda1 like I did wit Apr 28 03:15:31 h linux floppies in the olden days. I can't. Apr 28 03:16:55 The build was totally painless, but there is no information in the blog or sweetlilmre's README about actually booting the build. Apr 28 04:00:40 Aw Apr 28 04:01:01 beagleboard sdk failed due to the thread thing not being pushed yet Apr 28 04:01:06 * Tartarus goes to audit and fixup Apr 28 06:45:29 good morning Apr 28 06:45:37 Tartarus: I never tried vortex86sx + uclibc Apr 28 07:09:45 good morning Apr 28 07:26:11 What option in linux kernel 'menuconfig' do I have to set If I don't want to use kernel modules, but compile them into the kernel itself? Apr 28 07:26:49 Do I just have to disable 'Enable loadable module support'? Apr 28 07:29:15 tsjsieb: you have to set [*] in each module instead of [m] Apr 28 07:29:56 tsjsieb: 'Enable loadable module support' option allow you to load other modules later Apr 28 07:30:12 o.k. thnx :) Apr 28 08:21:32 morning all Apr 28 08:21:38 good morning Apr 28 08:21:55 * florian spots a pile of mails to answer but has no time :-( Apr 28 08:27:37 hey RP Apr 28 08:31:51 RP: hi Apr 28 08:32:32 RP: if you have that spare time...http://lists.arm.linux.org.uk/lurker/thread/20090427.120833.4ff23b55.en.html Apr 28 08:34:45 ant_work: Spare time? You are joking? ;-) Apr 28 08:34:52 hehe Apr 28 08:35:02 ant RP works 64 hours a day Apr 28 08:35:49 well, if you move fast enough the time will slow down around you :-) Apr 28 08:36:22 heh, RP get a superbike and code while bombing around track at over 10% of C Apr 28 08:40:56 XorA: Its an interesting idea :) Apr 28 08:41:19 XorA: The amount of damage the off road bike has done to me recently is making me wonder whether bikes are such a good idea though ;-) Apr 28 08:42:16 Well, its just making me think about being a little more careful... Apr 28 08:44:26 RP: was it the bike that damaged you? or the ground? Apr 28 08:44:42 RP: if the second then stop falling off :-D Apr 28 08:44:59 XorA: It would be the ground I guess :) Apr 28 08:45:23 The flying through the air bit was fine... Apr 28 08:46:32 time for a hitchhikers quote :-) Apr 28 09:07:56 RP: do you remember offhand if spitz has same ERASEBLOCKSIZE like akita ("0x20000") ? I suppose is the same on pxa270 machines Apr 28 09:08:26 just the nand was left untouched for spitz (until now) Apr 28 09:12:01 ant_work: I think its the same size eraseblock, yes Apr 28 09:12:21 cause we just have an override for ERASEBLOCKSIZE_akita = "0x20000" Apr 28 09:13:48 RP: btw I insanely found that patch to read old bootloader params :) Apr 28 09:14:08 http://www.katastrophos.net/zaurus/mirrors/piro.hopto.org/bootloader/patch-2.6.16.16-z.diff Apr 28 09:14:33 now I'll try to apply it to the linux-kexecboot kernel Apr 28 09:16:07 he..we missed that pre-kexec bootloader..still using a module to switch kernel... Apr 28 09:16:10 ant_work: I'm not sure thats the right patch ;-) Apr 28 09:16:23 CONFIG_SHARPSL_BOOTLDR_PARAMS ? Apr 28 09:16:35 we just need the damned mtdparts Apr 28 09:31:40 * XorA tries a build on FC11 Apr 28 09:37:26 can bitbake process delete the unpacked sources immediately after each completed task (like gentoo's emerge)? I am doing bitbake opie-image and it is all already over 5GB with only 1029 of 3339 done.. Apr 28 09:37:44 (hello :)) Apr 28 09:40:04 sledgeas: Look for the rm_work class Apr 28 09:41:28 thanks, RP; means something's missing on the getting started that i've followed (http://wiki.openembedded.net/index.php/Getting_started : (NOTE: kergoth says it will take about 4GB of disk space to build an opie or gpe image for one architecture) <-- absolutely does not hold for me being 1/3rd of way there :)) i may update that wiki in the end Apr 28 09:46:15 Here it's about 10GB, I think Apr 28 09:46:45 tsjsieb: you mean without INHERIT += "rm_work" ? Apr 28 09:47:02 without indeed Apr 28 09:47:45 phew :D Apr 28 09:48:50 but that's te complete working dir, including the openembedded git repository and the toolchain etc created with the openembedded recipes Apr 28 09:50:21 and the ./sources/ ? /) Apr 28 09:50:42 guys, is it safe on my Intel Core 2 Duo BB_NUMBER_THREADS = "2" && PARALLEL_MAKE="-j 2" or should I stick to one choice for opt performance? Apr 28 09:51:18 that's safe enought I think Apr 28 09:51:58 I only hope that it's not your desktop and you have to work on it, while compiling ;) Apr 28 09:52:02 yea, should make some benchmarkings myself Apr 28 09:52:04 :D Apr 28 09:52:18 lol no, i work on my workplace workstation currently Apr 28 09:52:29 leaving a more powerful box (a laptop lol) to do the bitbake job Apr 28 09:53:59 sledgeas: the download sources ar not included in the 10 gig, I have them in a sepparate dir, used by 2 different oe settups Apr 28 09:54:16 roger that Apr 28 09:54:52 sledgeas: "-j 3" (nr-of-cpu + 1) Apr 28 09:55:13 ant_work: what about BB_NUMBER_THREADS? Apr 28 09:55:25 I have 2 Apr 28 09:55:42 (lol surprisingly I had it -j 3, prolly from a gentoo's intuition I am runing on) Apr 28 09:55:55 gentoo here :-) Apr 28 09:56:29 download dir is about 1.5 GB here, but that could be smaller with a fresh one Apr 28 10:00:00 the wiki sure needs some adjustments, esp if one's (me:)) building opie-image from 0 having only 5GB+oe(+~1GB) dedicated :) Apr 28 10:01:41 sledgeas: with rm_work final build-size is approx 2.5gb for console-image + opie-image + x11-image. Sources extra. Apr 28 10:02:37 with 8GB RAM you can even build in /tmpfs Apr 28 10:02:50 with ccache too Apr 28 10:19:36 03Graeme Gregory  07org.openembedded.dev * rc21e4a50e3 10openembedded.git/recipes/unifdef/files/unifdef.c: Apr 28 10:19:36 recipes/unifdef/files/unifdef.c : more modern glibc expose getline from Apr 28 10:19:36 stdio.h rename the function so not to cause a clash Apr 28 10:40:58 03Graeme Gregory  07org.openembedded.dev * r9a67b35254 10openembedded.git/recipes/linux-libc-headers/ (files/unifdef.patch linux-libc-headers_2.6.23.bb): Apr 28 10:40:58 linux-libc-headers_2.6.23.bb : more modern glibc expose getline from Apr 28 10:40:58 stdio.h rename the function so not to cause a clash Apr 28 10:53:38 hi, I can't build the net-snmp_5.4.1.bb recipe which is in the openembedded/recipes/net-snmp directory. The build fails on configure task with this error: Apr 28 10:53:38 configure: error: cannot run /bin/sh /home/claudiu/oeDevel/tmpoe/work/i686-telefin-linux/net-snmp-5.4.1-r2/net-snmp-5.4.1/config.sub Apr 28 10:53:38 (the strange thing is that there is not config.sub file in the source directory) Apr 28 10:54:44 have someone experience with this package? Apr 28 11:42:24 03Koen Kooi  07org.openembedded.dev * rfa43579bcb 10openembedded.git/recipes/netkit-ftp/netkit-ftp_0.17.bb: netkit-ftp: Fixed GNU_HASH QA error. Apr 28 11:50:05 03Frans Meulenbroeks  07org.openembedded.dev * r8f5e54e67d 10openembedded.git/contrib/angstrom/build-feeds.sh: added netkit-ftp to angstrom feed Apr 28 11:58:40 03Robert Schuster  07stable/2009 * r9d392a418d 10openembedded.git/recipes/llvm/ (6 files): Apr 28 11:58:40 llvm-native 1.9: Removed. Apr 28 11:58:40 llvm-native 2.1: Removed. Apr 28 11:58:40 llvm-native 2.4: Removed. Apr 28 11:58:40 llvm-native cvs: Removed (llvm is not in a CVS repo anymore). Apr 28 11:58:42 llvm 2.0: Removed. Apr 28 11:58:44 llvm 2.4: Removed. Apr 28 11:58:46 03Robert Schuster  07stable/2009 * r25d182ad00 10openembedded.git/classes/java.bbclass: java.bbclass: Use force (-f) for ln (suggested by Marcin). Apr 28 11:58:51 03Robert Schuster  07stable/2009 * r05f14c7937 10openembedded.git/recipes/llvm/ (8 files in 3 dirs): (log message trimmed) Apr 28 11:58:54 llvm: Reworked to allow multiple installation of same version. Apr 28 11:58:56 - introduced common include file Apr 28 11:58:58 - provided implementation for llvm 2.5 Apr 28 11:59:00 - added wrapper script and its build recipe Apr 28 11:59:04 - the same for native side Apr 28 11:59:06 llvm2.5 2.5: New recipe Apr 28 11:59:08 03Robert Schuster  07stable/2009 * r18a9277dee 10openembedded.git/recipes/llvm/ (5 files in 2 dirs): Apr 28 11:59:11 llvm-gcc4-cross: Change dependency to llvm2.5-native. Apr 28 11:59:13 llvm-gcc: Dito. Apr 28 11:59:15 03Robert Schuster  07stable/2009 * r070c840053 10openembedded.git/conf/checksums.ini: checksums.ini: Removed llvm entries from 1.9 to 2.4. Apr 28 11:59:18 03Robert Schuster  07stable/2009 * rd3d2218c30 10openembedded.git/recipes/llvm/llvm.inc: llvm.inc: Declare dependency of llvm on llvm-native Apr 28 11:59:23 03Robert Schuster  07stable/2009 * rda7489f755 10openembedded.git/recipes/xorg-lib/ (libxt/pr10970-header-fix.patch libxt_1.0.4.bb libxt_1.0.5.bb): Apr 28 11:59:26 libxt 1.0.4: Added official patch to fix X.org PR10970. Apr 28 11:59:28 libxt 1.0.5: Dito. Apr 28 11:59:30 libxt-native 1.0.5: Dito. Apr 28 12:00:37 03Robert Schuster  07stable/2009 * r504781534d 10openembedded.git/recipes/lesstif/lesstif_0.95.0.bb: lesstif 0.95: Actually install something into staging. Apr 28 12:01:31 seen ~rschuster Apr 28 12:01:42 oops Apr 28 12:01:51 ~seen rschuster Apr 28 12:01:55 rschuster was last seen on IRC in channel #oe, 30d 19h 57m 36s ago, saying: 'you need to apply a trick with EXTRA_OEMAKE (the value for normal make) to get it work'. Apr 28 12:02:08 seen ~theboemian Apr 28 12:02:15 ~seen theboemian Apr 28 12:02:16 Gnutoo: i haven't seen 'theboemian' Apr 28 12:02:31 thebohemian, hi Apr 28 12:03:38 thebohemian, I've a bug in boost-38 that is not related to the assembly part that I added... Apr 28 12:03:54 thebohemian, it prevent wesnoth from running on armv4 Apr 28 12:04:05 I've the full stack trace Apr 28 12:04:08 but I must: Apr 28 12:04:18 *try with boost from trunk Apr 28 12:04:22 *bugreport Apr 28 12:04:53 so should I commit wesnoth and add DEFAULT_PREFERENCE_armv4 = "-1" and do the same for armv5 and armv4t? Apr 28 12:05:11 s/commit/send to the mailing list/ Apr 28 12:16:45 Gnutoo: sounds ok to me Apr 28 12:16:55 Gnutoo: how difficult is this boost 38 issue? Apr 28 12:17:38 I will bugreport when I'll have the time so it's not me that will resolve the problem Apr 28 12:17:46 I have backtrace if you want Apr 28 12:18:02 http://gnutoo.homelinux.org/wesnoth/backtrace_0002.txt Apr 28 12:18:09 Gnutoo: they have boost 38 in debian Apr 28 12:18:12 the main issue is time Apr 28 12:18:19 wow Apr 28 12:18:25 so they may have patches for it Apr 28 12:18:40 Gnutoo: yeah, that is what I would suggest Apr 28 12:18:41 the problem is that wesnoth and boost-38 compiles fine Apr 28 12:19:12 only that wesnoth segfault when launching the tutorial,a campaing etc... Apr 28 12:19:40 Debian's wesnoth is against boost 1.37. not sure if this is an important thing Apr 28 12:19:46 also I've a big problem with 320x240 screens...they added some great graphic things but put assertions Apr 28 12:19:49 ok Apr 28 12:20:23 the assertion make the game exit on 320x240 screen Apr 28 12:20:27 Gnutoo: btw. you cannot have DEFAULT_PREFERENCE_armv4t Apr 28 12:20:38 ouch...so what should I do? Apr 28 12:20:42 Gnutoo: only "_arm" Apr 28 12:20:43 @python? Apr 28 12:20:45 ok Apr 28 12:21:13 Gnutoo: using _armv4t etc. is strongly discouraged (we once had this in llvm) Apr 28 12:21:30 if armv4 is defined => default preference = -1 ? Apr 28 12:22:20 Gnutoo: I mean libffi. It is actually still there (not good) Apr 28 12:22:21 03Stanislav Brabec  07stable/2009 * r450798417c 10openembedded.git/recipes/unzip/unzip_552.bb: Apr 28 12:22:21 unzip: GNU_HASH QA fix. Apr 28 12:22:21 Signed-off-by: Koen Kooi Apr 28 12:22:21 Acked-by: Denys Dmytriyenko Apr 28 12:22:22 03Stanislav Brabec  07stable/2009 * rf3ffb7a22c 10openembedded.git/recipes/man/man_1.5p.bb: Apr 28 12:22:24 man: GNU_HASH QA fix. Apr 28 12:22:26 Signed-off-by: Koen Kooi Apr 28 12:22:30 Acked-by: Denys Dmytriyenko Apr 28 12:22:34 03Stanislav Brabec  07stable/2009 * r8d27f730a1 10openembedded.git/recipes/pxaregs/pxaregs_1.14.bb: Apr 28 12:22:36 ok Apr 28 12:22:37 pxaregs: GNU_HASH QA fix. Apr 28 12:22:39 Signed-off-by: Koen Kooi Apr 28 12:22:41 Acked-by: Denys Dmytriyenko Apr 28 12:22:41 Gnutoo: the bug affects all arm arches, right? Apr 28 12:22:43 03Koen Kooi  07stable/2009 * rc4cd2af638 10openembedded.git/recipes/netkit-ftp/netkit-ftp_0.17.bb: Apr 28 12:22:46 netkit-ftp: Fixed GNU_HASH QA error. Apr 28 12:22:50 Signed-off-by: Koen Kooi Apr 28 12:22:51 no arm < armv6 Apr 28 12:22:52 Acked-by: Philip Balister Apr 28 12:23:14 Gnutoo: ok Apr 28 12:23:28 Gnutoo: another atomic compare and swap issue? Apr 28 12:23:31 I haven't tested on armv5 Apr 28 12:23:46 don't know...did you look at the stacktrace Apr 28 12:23:48 ? Apr 28 12:25:17 by the way here's the patch: http://www.pastebin.ca/1406020 Apr 28 12:25:43 but if I remember well return must be changed Apr 28 12:25:51 my C++ is definitely not good enough to understand what happens there :) Apr 28 12:26:26 I suggest looking at this in the debugger. my gut feeling is that it may have something to do with some static variable not being defined Apr 28 12:26:34 the important part is the assembly Apr 28 12:27:13 I don't know C++ either...I learn the minimal part necessary for understanding what was going on...only for wesnoth/boost Apr 28 12:27:22 s/learn/learnt/ Apr 28 12:27:38 ah ok thanks Apr 28 12:28:18 -> http://gnutoo.homelinux.org/wesnoth/backtrace_0002.txt: no assembly here Apr 28 12:28:20 (the assembly part must also be adapted for armv4...changing the jump for a non-thumb thing) Apr 28 12:28:29 ah in the patch ... Apr 28 12:28:33 yes Apr 28 12:28:56 I'll sum up what I did with some help Apr 28 12:28:57 : Apr 28 12:29:11 *made wesnoth compile with the patch for boost Apr 28 12:29:21 (I was helped a lot for the patch) Apr 28 12:29:42 then wesnoth ran but was unplayable...you had only the main menu that worked Apr 28 12:30:45 the backtrace was obtained running wesnoth and clicking on the menu's tutorial button...it loaded...you even saw a first image of the game...then segfaulted Apr 28 12:30:56 before loading the lower image of the character Apr 28 12:31:10 the assembly explanation is at: Apr 28 12:31:19 http://lwn.net/Articles/314144/ Apr 28 12:31:33 at: Posted Jan 12, 2009 19:58 UTC (Mon) by npitre (subscriber, #5680) [Link] Apr 28 12:34:47 Gnutoo: I would rather not use hardcoded assembly Apr 28 12:35:26 Gnutoo: since the assembly is tied to the ABI in use Apr 28 12:35:40 and there are two for ARM EABI and OABI Apr 28 12:36:20 you can do what they do in C/C++ as well. and it should solve your problem as well Apr 28 12:37:15 ok Apr 28 12:37:27 thanks a lot Apr 28 12:37:47 and I've even an arm oabi device...lol Apr 28 12:37:51 is there a way to fight my firewall i.e.: how could I redirect all requests to ftp://.*/([a-zA-Z_-0-9\.].tar.gz) into a mirror e.g. http://www.angstrom-distribution.org/unstable/sources/\1 ? Apr 28 12:37:56 (kernel 2.4) Apr 28 12:38:11 thanks a lot Apr 28 12:38:34 thebohemian, but in C...how is the atomicity guaranteed? Apr 28 12:38:43 s/C/C\/C++/ Apr 28 12:39:25 or maybe I should do one for arm-eabi and then one for arm oabi Apr 28 12:41:04 Gnutoo: http://www.pastebin.ca/1406031 Apr 28 12:41:15 Gnutoo: the atomicity is guaranteed by the kernel Apr 28 12:41:30 ok thanks a lot Apr 28 12:41:37 Gnutoo: it is the same as the assembly except that you let the compiler create the machine code Apr 28 12:41:48 ok thanks a lot Apr 28 12:41:52 wow Apr 28 12:42:05 should I remove the armv7 code? Apr 28 12:42:21 just call the the arm_add_and_fetch function instead of __sync_add_and_fetch Apr 28 12:42:27 you can leave armv7 stuff in Apr 28 12:42:30 ok thanks Apr 28 12:42:46 it will be faster on this arch since it does not need the kernel's help Apr 28 12:43:16 basically arm_add_and_fetch is what newer GCCs provide behing __sync_add_and_fetch Apr 28 12:43:26 s/behing/behind/ Apr 28 12:44:30 the kernel helper is usefull for anything below armv6 btw. Apr 28 12:45:40 so it should look like: Apr 28 12:45:40 ifdef ARM Apr 28 12:45:40 ifdef ARMv6 or ARMv7 Apr 28 12:45:40 assembly Apr 28 12:45:40 else Apr 28 12:45:41 arm_add_and_fetch(val, 1) Apr 28 12:45:43 endif Apr 28 12:45:45 else Apr 28 12:45:47 __sync_add_and_fetch(val, 1) Apr 28 12:45:49 endif Apr 28 12:46:06 (and -1 for operator--, of course) Apr 28 12:47:18 thanks a lot that was what I had in mind Apr 28 12:51:10 03Koen Kooi  07org.openembedded.dev * rf37fe230b7 10openembedded.git/recipes/ekiga/ekiga_git.bb: ekiga: add git version Apr 28 12:54:30 is there a way to swap the order of mirrors that bitbakes fetches from and use only http-based ones? Apr 28 12:54:41 (opie-image, angstrom) Apr 28 12:55:35 morning Apr 28 12:57:21 gooday Apr 28 13:03:06 hrw: good afternoon Apr 28 13:04:25 i want to compile a kernel that is able to execute the "old" data format prior the eabi, can this be done with an eabi build environment, ? changeing the kernel options has no result (:-(( Apr 28 13:05:11 where did you change the kernel options? Apr 28 13:06:10 I do use oABI support in a kernel, compiled in devshell Apr 28 13:06:18 in the kernel . ( make config , export ARCH=arm ; change to the kernel source , configure it, compile it , still not able to execute debian arm binary ) Apr 28 13:08:29 i have a full featured debian installation, were running nice since the last years. with the migration to kexec all is lost .. Apr 28 13:09:22 I have no idea what's going wrong in your situation, but I am able to use my kernel, created within OE, with an debian 4.0 arm (oABI) rootfs, so I guess it should be possible for you too Apr 28 13:09:30 kexec cannot start old kernel versions . ( no commandline options were passed ) Apr 28 13:09:59 Angstrom turns off CONFIG_OABI_COMPAT for all kernels except gta01 Apr 28 13:11:38 XorA: i have tried with CONFIG_OABI_COMPAT, and without it , all combinations of this stuff , allways the same ... Illegal instruction . Apr 28 13:11:59 pwgen: building in OE with Angstrom though? Apr 28 13:12:43 XorA: it doesn't do that when you compile a kernel manualy within devshell, does it? Apr 28 13:13:08 tsjsieb: your .config will still be mangled by linux.inc Apr 28 13:13:21 tsjsieb: unless you fix it up afterwards Apr 28 13:13:24 the only way getting a kernel that can execute someting is compiling it native on the zaurus. -> chicken egg, Apr 28 13:14:14 o.k., thnx, that's what I'm doing for the moment :) Apr 28 13:14:37 i can install some old pdaxrom image, and with that i can chroot to the system, but this will allways overwrite my kexecboot image Apr 28 13:16:04 i see. ever change a running system .. if you dont, then comes the moment you need to erase all .. (:-(( Apr 28 13:19:42 pwgen: maybe you hit a genuine kernel bug Apr 28 13:21:11 (:-)) maybe a complete bunch of it .. ( for example . kexec starts my old 2.6.17, but cannot pass the kernel-cmdline, so it trys allways to boot from mtdblock2 ) Apr 28 13:27:46 03Gregoire Gentil  07org.openembedded.dev * r015cd5d427 10openembedded.git/recipes/gstreamer/ (2 files in 2 dirs): gst-omapfb: add recipe to build gst sink for the omapfb overlay Apr 28 13:28:02 florian, any word on linuxtag? Apr 28 13:36:29 pwgen: have you checked the /boot/kernel-cmdline? Apr 28 13:38:37 hello is there any mantainer here? Apr 28 13:38:51 quick everyone hide Apr 28 13:39:04 :) Apr 28 13:39:09 * Crofton|work ducks Apr 28 13:39:14 maintainer of what? Apr 28 13:39:34 i need a patch to be acked Apr 28 13:39:54 i sent it in the openembedded-devel@lists.openembedded.org list Apr 28 13:40:19 so i'd like to talk about it with some mantainer who can ack it Apr 28 13:40:28 which patch? Apr 28 13:40:35 zenity package Apr 28 13:41:18 http://patchwork.openembedded.org/patch/416/ Apr 28 13:41:41 hmm when do our python coders update bitbake for all the deprecation in python 2.6? Apr 28 13:41:50 yes that one Apr 28 13:42:01 : i can see in the serial console what commandline parameter are used .. i am 100% sure, they are not the cmdlineparameter kexecboot used Apr 28 13:42:58 @crofton well there were other attachments in the mail Apr 28 13:43:13 hmm Apr 28 13:43:26 hariseldon78_: non git-diff patches tend to get ignored Apr 28 13:43:33 it does not likk like we can easily apply it Apr 28 13:43:34 thebohemian: ping Apr 28 13:43:50 hariseldon78_, can you make it with git diif Apr 28 13:44:00 or some other way it can easily be applied? Apr 28 13:44:14 ok i will try Apr 28 13:44:17 thanks Apr 28 13:44:27 how much git do you know? Apr 28 13:44:33 hrw: pong Apr 28 13:44:43 thebohemian: your llvm changes in stable/2009... Apr 28 13:44:48 should i checkout the last version from git and get a patch from that? Apr 28 13:44:56 hrw: uh oh. something wrong? Apr 28 13:44:58 yeah Apr 28 13:44:58 not much but i'm learning Apr 28 13:44:59 thebohemian: they lack any ack Apr 28 13:45:03 ok Apr 28 13:45:20 hariseldon78_: you need to get latest OE and prepare your patches against that Apr 28 13:45:36 perfect Apr 28 13:45:37 hrw: ouch. what do I need to do? Apr 28 13:45:39 i'll do that Apr 28 13:45:40 thebohemian: stable/2009 require at least one Acked-by on each changeset Apr 28 13:45:43 hariseldon78_: basically so we can see the .bb file changes and the new files you added Apr 28 13:45:44 thank you Apr 28 13:45:45 I'm going to mark that patch in patchwork as something so they disappear from the list of stuff we need to work on Apr 28 13:45:55 yes please Apr 28 13:45:56 hrw: sorry, was not aware of this Apr 28 13:46:00 i'll resend it Apr 28 13:46:02 thebohemian: http://wiki.openembedded.net/index.php/Stable#How_to_share_changes Apr 28 13:46:23 I marked it as rejected, but don't take that the wrong way :) Apr 28 13:46:34 yes no problem Apr 28 13:46:38 ah one thing Apr 28 13:46:43 thebohemian: I would revert them and wait for acked versions Apr 28 13:46:52 in the bb recipe i used inherit autotools Apr 28 13:47:05 http://patchwork.openembedded.org/patch/308/ Apr 28 13:47:12 what about this patch? Apr 28 13:47:16 thebohemian: and by 'them' I mean all your changes pushed today Apr 28 13:47:19 did it get applied? Apr 28 13:47:24 even if the package has some gnome dependencies and get downloaded from the gnome ftp Apr 28 13:47:42 hrw: yep Apr 28 13:47:43 Crofton|work: what is your opinion? Apr 28 13:47:46 yes that one is applied Apr 28 13:47:53 ok Apr 28 13:48:08 hey hrw, do you ever have randomish failures w/ your buildbot? Apr 28 13:48:19 Tartarus: mostly git pull problems Apr 28 13:48:20 i think gnome.bbclass should be used but when i try the build fails strangely Apr 28 13:48:27 Tartarus: sometimes udev unpack problem Apr 28 13:48:38 hmm Apr 28 13:48:49 some "/usr/include" paqths get into the makefiles and make the crosscompiler fail Apr 28 13:48:55 hrw, urg Apr 28 13:49:14 binutils-cross failed on almost every target for me overnight Apr 28 13:49:15 do the patches acks and they just are listed in the commit message? Apr 28 13:49:17 except once when ipkg-native did Apr 28 13:49:21 so autotools is the best i can do by now, i hope to fix it for some other versions Apr 28 13:50:01 ok, i'll explain it in the patch mail Apr 28 13:50:03 bye Apr 28 13:50:05 hariseldon78_, make a patch we can look at, and people can comment Apr 28 13:50:12 thanks! Apr 28 13:50:15 sure, thanks Apr 28 13:51:36 hrw: sure Apr 28 13:51:42 hrw, I'd like to see the commits without acks reverted and pushed again with acks Apr 28 13:51:46 pwgen: look here to have an idea... http://lkml.org/lkml/2007/10/24/662 Apr 28 13:52:00 I know that koen hates it when people remove checksum entries .... Apr 28 13:52:08 so I would leave that one out Apr 28 13:52:24 hrw: I used 'git cherry-pick' to get these from .dev. when I revert some the patches to .stable can I use 'cherry-pick' again? Apr 28 13:52:39 s/some// Apr 28 13:52:45 thebohemian, not sure, but i think we will find out :) Apr 28 13:53:42 and thanks for working with us Apr 28 13:54:10 just use git-reset to rewind and add acks to commits, then git-format-patch to extract the patches Apr 28 13:54:28 revert all the wrong ones, re-apply the patches now with acks Apr 28 13:58:15 03Marcin Juszkiewicz  07stable/2009 * re264996efb 10openembedded.git/ (19 files in 9 dirs): Apr 28 13:58:15 Reverted all commits done by Robert Schuster. Apr 28 13:58:15 Each change to stable/2009 branch needs to be reviewed and acked by Apr 28 13:58:15 stable branch maintainers. Apr 28 13:58:15 Signed-off-by: Marcin Juszkiewicz Apr 28 13:58:44 ant_work: the kexec sems a little bit unstable, on OM GTA02 its completly broken with 2.6.26 , i think i need for any kernel i use a special selected kexec Apr 28 13:58:48 thebohemian: git cherry-pick, then rebase, git format-patch -n -s, git send-email Apr 28 13:59:07 thebohemian: we review, ack and then you grab them from patchwork and push Apr 28 14:06:01 ant_work: http://pastebin.com/m432a624c Line 5. ignoring tag 000. this is the result when starting 2.6.20 with kexec build with/for 2.6.26 Apr 28 14:08:06 hrw: ok Apr 28 14:08:42 Fedora Core 11 Beta, ext4 filesystem NOTE: build 200904281337: completed Apr 28 14:09:09 running fedora betas you masochist Apr 28 14:09:19 XorA: Didn't know you ran Fedora. Apr 28 14:09:21 haha yes thats evil Apr 28 14:09:26 pwgen: Ignoring unrecognised tag 0x00000000 Apr 28 14:09:30 Crofton|work: I thought I would get ahead of the curve on this one Apr 28 14:09:31 more evil then running debian unstable Apr 28 14:09:36 (and I didn't think that ext4 was stable enough for that!) Apr 28 14:09:39 pwgen: this is broken u-boot iirc Apr 28 14:09:39 mwester: I got bored of Ubuggeredtu Apr 28 14:09:40 XorA, how do they handle libusb1/libusb-compat? Apr 28 14:10:01 mwester only when you plug powercable Apr 28 14:10:04 * Crofton|work is surrounded by various Feora machines Apr 28 14:10:09 woglinde: hehe! Apr 28 14:10:53 * mwester downloaded fc11-beta; maybe he should load it up on something around here.... Apr 28 14:11:34 I'm going to upgrade my older installs to 11 when it is releaseed Apr 28 14:11:42 Crofton|work: they seem to have both libusb and libusb1 Apr 28 14:11:47 ant_work: at this time uboot has started an kexec-intrd image, this bootmessage is the output after kexec has executed the zImage from /dev/hda1/boot Apr 28 14:11:58 so they did not use libusb-compat Apr 28 14:12:02 crap Apr 28 14:12:13 I was hoping I could test some libusb-compat stuff on it Apr 28 14:12:15 pwgen: kernel before 2.6.23-something deliberately ignored atags/cmdline passed by bootloader Apr 28 14:12:41 linux-rp Apr 28 14:12:51 XorA: did you disable locale generation on FC11? Apr 28 14:13:03 mwester: this was beagle so that doesnt work anyway Apr 28 14:13:11 Ah, ok. Apr 28 14:13:28 ant_work: i backup home and configurations an try to build some similar with the zubuntu ( i see no chance to update the old debian installaion ) Apr 28 14:13:33 * XorA is trying to ignore old crappy are devices :-) Apr 28 14:13:46 pwgen: by the way you have 'strange' mtdparts Apr 28 14:13:51 I forgot to disable locales on my FC10 machine the other day, but the build completed anyway -- so I am wondering if finally the qemu/fedora problems have been magically resolved. Apr 28 14:15:07 ant_work: i use the same setup on an c-1000 with 'strange' broken nand (:-)) Apr 28 14:15:17 BTW the Beta disk was a cockup release the kernel will oops until upgraded :-) Apr 28 14:15:39 pwgen: I don't see Root :/ Apr 28 14:16:18 pwgen: my old pastebins have disappeared.. Apr 28 14:17:04 this one was booting without mtd compiled in iirc http://ja.pastebin.ca/1173295?srch=noop Apr 28 14:17:30 kexecboot should pass the mtdparts... Apr 28 14:17:47 these are not in your pastebin... Apr 28 14:17:51 after the other device has killed its nand i tried to avoid useing it and use the internal microdrive ( maybe i can replace it without soldering ) Apr 28 14:18:07 hi woglinde Apr 28 14:18:21 hi mrdata Apr 28 14:18:38 I will start a test build today Apr 28 14:18:41 ant_work: kexec is passing nothing to this kernel .. thats on of my problems Apr 28 14:19:10 well, kexec passes /boot/kernel-cmdline and appends mtdparts Apr 28 14:19:24 woglinde: fine, i hope it will be okay ;-) Apr 28 14:19:47 pwgen: actually cmdline is in /boot/boot.cfg -> APPEND Apr 28 14:20:18 with latest kexecboot from cfgfiles branch Apr 28 14:21:37 pwgen: found the pastebin...http://ja.pastebin.ca/1173317 Apr 28 14:21:38 .. i kown what kexec is doing , but the kernel is ignoring it ( kexec -l zImage --command-line="ignore all my options" will do absolutly nothing with 2.6.20 or .17 ) Apr 28 14:22:05 pwgen: look at the a.n. log... Apr 28 14:23:47 jo mickeyl Apr 28 14:23:51 morning Apr 28 14:25:02 by the way , who has written the tskeys app ? Apr 28 14:25:20 RP wrote tskeys I think Apr 28 14:25:21 you mean tslib? Apr 28 14:25:29 what is the exact procedure to submit a stable patch? Apr 28 14:25:31 part of zaurusd Apr 28 14:25:44 http://wiki.openembedded.net/index.php/Stable doesn't say how to send the mail Apr 28 14:25:44 mckoan put on the dev ml Apr 28 14:25:46 and get acks Apr 28 14:26:00 do I have to attch the patch file? Apr 28 14:26:05 yes Apr 28 14:26:09 i have always al load of 1,0 generated by tskeys Apr 28 14:26:13 or include its code? Apr 28 14:26:15 patchwork will pick it up Apr 28 14:26:29 read man git-fromat-patch Apr 28 14:26:31 format Apr 28 14:26:39 pwgen: something might fire a HUP or ENOAGAIN on the fds Apr 28 14:26:51 pwgen: 2.6.20 and 2.6.17 probably lack atag's patch. cmdline is compiled-in (ignore bogus bootloader params) Apr 28 14:26:52 same result but different method, what is appropriate? Apr 28 14:26:52 pwgen: add some debug statements where it reads from the input devices Apr 28 14:27:03 mckoan???? Apr 28 14:27:19 mckoan get familiar with git format-patch Apr 28 14:27:28 send patches do the -dev ml Apr 28 14:27:29 woglinde: I already did it Apr 28 14:27:31 get acks Apr 28 14:27:35 und then sumbit Apr 28 14:27:38 woglinde: I know Apr 28 14:27:40 submit to stable Apr 28 14:27:55 if you dont have acess pester the ml with patches Apr 28 14:28:01 and get rw access Apr 28 14:28:13 if the patches have some quality standards Apr 28 14:28:41 woglinde: I also see that almost all patched have "Signed-off-by", is this added by hand? Apr 28 14:29:07 koen wrote an easy howto to the dev ml Apr 28 14:29:12 ant_work: the reanimation of the old image took now 4 days . its time now to bury it to dev/null .. Apr 28 14:29:27 woglinde: do you mean this? http://wiki.openembedded.net/index.php/Stable Apr 28 14:29:33 pwgen: :/ Apr 28 14:31:13 mckoan thats describe the process Apr 28 14:32:42 woglinde: http://wiki.openembedded.net/index.php/Commit_Policy Apr 28 14:32:57 I'll try doing my best Apr 28 14:35:32 ant_work: maybe i can reactivate it inside an nfs chroot in future , to extract all configurations , lets have a look at zubuntu and hope that they provide the same ammount of packages Apr 28 14:35:54 mckoan hm some infos on patchwork are here Apr 28 14:35:56 http://article.gmane.org/gmane.comp.handhelds.openembedded/22445/match=patchwork Apr 28 14:36:53 woglinde: what patchwork is? Apr 28 14:37:15 http://patchwork.openembedded.org/project/openembedded/list/ Apr 28 14:39:17 woglinde: but a "how to send a patch and where" exists? Apr 28 14:39:22 i have the suspect, that with the 2.6.26 the g_ether thingie is also broken for the zaurus .. Apr 28 14:39:39 mckoan send the patch to -dev ml and it will end up in patchwork Apr 28 14:39:54 woglinde: automatically? Apr 28 14:40:00 YESSSSSSSSSSSSSSSSS Apr 28 14:40:06 *sig* Apr 28 14:40:55 woglinde: patch mail have to be formatted is any way? Apr 28 14:41:04 s/is/in Apr 28 14:41:27 mckoan which point is now unclear for you? Apr 28 14:41:45 I dont what to say the same stuff the tenth time Apr 28 14:43:15 mickeyl: better soltution for tskeys: rm /usr/bin/tskeys Apr 28 14:43:17 mckoan: "git format-patch" patches Apr 28 14:43:42 hrw I did git format-patch -1 Apr 28 14:43:43 hrw hm how does the sign ack goes in? Apr 28 14:43:49 hrw are the typed by hand? Apr 28 14:43:49 pwgen: well, to defend it... it's old Apr 28 14:44:19 woglinde: is typed by hand, I asked before doing that Apr 28 14:44:47 mckoan *sigh* so you know all the steps you have to do Apr 28 14:44:54 mckoan you asked them again Apr 28 14:44:55 * Tartarus begs for people to comment on / review his recent RFCs so he can send a few more out Apr 28 14:44:57 woglinde: contrib/patchwork/ in .dev has a script for that Apr 28 14:45:12 mickey: that was the suspect i had .. not finished . ( i never missed the 5 function pads at the touchscreen , they are as buttons on the keyboard ) ) Apr 28 14:46:02 my first question was "what is the exact procedure to submit a stable patch?" Apr 28 14:46:47 but should be "what is the exact procedure to generate and submit a stable patch?" Apr 28 14:46:58 it the power consumption for an arm cpu the same for doing idle task or calculation something ? Apr 28 14:48:45 03Koen Kooi  07org.openembedded.dev * r4c992cda6f 10openembedded.git/recipes/zaurusd/ (2 files in 2 dirs): zaurusd: point tskeys to the real ts device instead hardcoding it to event1 Apr 28 14:52:15 mickeyl: did you think about upstart ? Apr 28 14:53:10 hrw, woglinde, I don't understand if the patch I submitted is ok or not Apr 28 14:53:32 mckoan test it Apr 28 14:53:43 git format-patch takes care Apr 28 14:54:09 pwgen: not much lately. i will probably try to incorporate it in my OE fork Apr 28 14:54:11 woglinde: what do you mean with "test it" Apr 28 14:54:16 and then the reviewers will tell you what is wrong in the sense of oe Apr 28 14:54:27 mckoan test your patch local Apr 28 14:54:40 with as many combination you think it can break Apr 28 14:54:51 woglinde: my local copy has already the patch Apr 28 14:55:06 woglinde: do you mean in another local copy? Apr 28 14:56:05 mwester: NOTE: generating locale es_NI (UTF-8) Apr 28 14:56:06 NOTE: generating locale gv_GB (UTF-8) Apr 28 14:56:35 mckoan *sigh* Apr 28 14:56:56 mckoan what you know about software engineering? Apr 28 14:57:00 gv_GB? Interesting... Apr 28 14:57:25 woglinde: I never used patchwork and git before Apr 28 14:57:45 broonie: is that Gaelic then? Apr 28 14:57:49 woglinde: I never submitted patches directly to OE ML Apr 28 14:57:49 mckoan this has nothing to do with git and patchwork Apr 28 14:57:59 mckoan test if it breaks Apr 28 14:58:07 then use git format-patch Apr 28 14:58:13 send output to the ml Apr 28 14:58:17 get the reviews Apr 28 14:58:18 woglinde: yes it has, I understand the problem is mainly related to policies Apr 28 14:58:23 whats so jhard to understand? Apr 28 14:58:33 XorA: Isle of Man gaelic AFAIR. Apr 28 14:58:45 are you analyzing easy stuff you can try out in 10 seconds for month? Apr 28 14:58:51 woglinde: I'm really frustrated my this, please apologize me Apr 28 14:59:05 mckoan sometimes just try Apr 28 14:59:13 when it was wrong you will learn from it Apr 28 14:59:15 woglinde: I know how to generate a patch with git format-patch Apr 28 14:59:24 woglinde: but OE policies ask more Apr 28 14:59:25 but discussing it for months and hours will you get nowhere Apr 28 14:59:34 woglinde: ok Apr 28 14:59:45 woglinde: so is my patch ok or not? Apr 28 14:59:54 seems I understand now why the others dont talk to you Apr 28 15:00:54 woglinde: a simple answer would help Apr 28 15:01:20 I gave you the simple anwser 5 times Apr 28 15:01:39 I can try submitting other patches, but I don't want to flood ML with wron ones Apr 28 15:02:03 mckoan pleas try git format-patch Apr 28 15:02:09 it will not harm animals Apr 28 15:02:25 jo kergoth Apr 28 15:02:30 mckoan: This is the free software community; people will be delighted to tell you if you get stuff wrong. Apr 28 15:02:39 hehe! Apr 28 15:02:45 * mickeyl coughs Apr 28 15:03:02 I could wine or laugh Apr 28 15:03:05 hey Apr 28 15:03:37 woglinde: I thank you for your answers, tell me if I'm annoying you Apr 28 15:04:13 mckoan I have kids Apr 28 15:04:21 I can take a lot of anger Apr 28 15:04:31 ups only one kid Apr 28 15:05:38 woglinde: I never take anger with mine Apr 28 15:05:50 :_D Apr 28 15:06:04 mckoan so they do always what you tell them Apr 28 15:06:39 * likewise_ suspects mckoan applies policies :-) Apr 28 15:06:50 * likewise_ doesn't have kids though Apr 28 15:07:10 BTW :-) Apr 28 15:07:33 if I do git format-patch -1 again what should happen ? Apr 28 15:07:54 just do it Apr 28 15:07:57 you will see Apr 28 15:08:12 ok done Apr 28 15:09:16 * mckoan submitting again the patch 'as-is' without signed-by Apr 28 15:10:16 mckoan: "-s" added to git format-patch adds signed-by line Apr 28 15:10:42 hrw: thank you, this is useful answer Apr 28 15:12:21 okay I have to bye Apr 28 15:12:27 till later Apr 28 15:12:28 cya woglinde Apr 28 15:13:04 woglinde: thank you Apr 28 15:14:10 mickeyl: good afternoon Apr 28 15:14:20 greetings pb_, how are things? Apr 28 15:14:32 yay! my buildbot is finally green Apr 28 15:14:58 nice :) Apr 28 15:15:18 hrw, so you can review the wireless-tools change for stable now then? ;) Apr 28 15:15:49 mickeyl: pretty good, preparing for my paternity leave at the moment :-} Apr 28 15:16:20 ah, excellent! Apr 28 15:16:33 will you take the whole time or share? Apr 28 15:17:55 not sure yet, going to wait and see how it works out. Apr 28 15:17:58 I haven't done this before :-) Apr 28 15:18:29 also just catching up on my mailing list mails, I was sorry to see your note from the other day. Apr 28 15:18:32 pb_: get hrw to write a recipe for it Apr 28 15:18:36 heh Apr 28 15:18:37 Tartarus: yes. which patch it is in patchwork? Apr 28 15:19:13 can we automate acking patches, like in "./ack.sh 334" somehow? Apr 28 15:19:36 http://patchwork.openembedded.org/bundle/trini/trini-stable-2009-sdk-updates/ has it :) Apr 28 15:20:06 thanks, just read your reply. yeah, it finally came to the point where i'm too weak to continue fighting,so i had to capitulate. i don't really know how to proceed now, but i know i can't just continue comitting to OE as if nothing were, so i have to find something else. Apr 28 15:20:43 i do like the idea of a community elected team very much Apr 28 15:21:45 hrw, I hate how reasoning for changes seems to get stripped out of git send-email stuff, see http://cgit.openembedded.net/cgit.cgi?url=openembedded/commit/&id=28d0f54e301b5063eaf3587c57fd5422ae7fb2b2 for the how/why Apr 28 15:23:37 mckoan: your patch fails first test, its not in .dev Apr 28 15:25:01 Tartarus: 17:24 hrw@home:stable$ scripts/build-test-gu.sh wireless-tools Apr 28 15:25:11 XorA: now that I know how to prepare it I can submit more :-) Apr 28 15:25:20 HI Apr 28 15:25:58 I have done a "bitbake sdlquake" this morning for the target omap3evm and the angstrom distribution Apr 28 15:26:17 My build is a sucess :) Apr 28 15:26:52 mckoan: BTW why did you do an exact copy of Angstrom? Apr 28 15:27:05 But now I am looking for a eay way to run the quake application on My filesystem Apr 28 15:27:24 And I am not sure howto do this Apr 28 15:27:53 * mckoan doesn't know why others shouldn't talk to me as woglinde said... I like to help each time I can Apr 28 15:27:53 mickeyl: yah, I've always been a bit reluctant to propose elections in the past because (a) it seems a bit too much like the kind of bureaucratic navel-gazing that we all loathe so much, and (b) it's not entirely clear to me how you would define the set of people eligible to vote, since oe doesn't really have a formal membership as such. But I guess one could just use the list of authorised git keys or some such as the electoral roll. Apr 28 15:27:56 I have seen that normallt some .ipk or .deb should be created but I couldn't find any ? Apr 28 15:28:23 XorA: It was already likely angstrom, this is a starting point Apr 28 15:28:32 * mwester would favor some formalization of who can "vote", who can "ack/nack" patches, etc. Apr 28 15:28:45 yeah, that would be nice. Apr 28 15:28:53 hrw, were you (cross)compiling toolchains? Apr 28 15:28:57 I've never known whether I am entitled to ack or nack patches for example Apr 28 15:29:06 Exactly. Apr 28 15:29:10 very true Apr 28 15:29:11 So I abstain, mostly. Apr 28 15:30:00 XorA: I thought that for this work I would have people helping me, but preferred to do something else Apr 28 15:30:52 mckoan: normally just requiring angstrom.conf then changing the DISTRO = "blah" works well and means you dont have to chase Angstrom Apr 28 15:32:08 XorA: BTW the current kaeilos.conf fails with qemux86 so I *have to* patch it Apr 28 15:37:35 i still don't think a "core team" is viable, voted for or not, myself. i think we need some specific roles filled by specific folks, so you know who's responsible for it, not a fuzzy group of individuals. make person A responsible for CM, person B responsible for overall architectural direction for the project, etc Apr 28 15:39:24 well, I think a core team would be fine so long as it actually took decisions and was seen to do so. gcc, for example, seems to function just fine with a steering committee of something like ten people. Apr 28 15:39:27 kergoth: agree Apr 28 15:39:47 * hrw wonder what OE e.V. status has about it Apr 28 15:40:08 Tartarus: wireless-tools builds fine so far Apr 28 15:40:33 true. itd help if announcements were made, and there needs to be a mechanism of resolving conflicts. today it seems like when cor emembers disagree, either nothing ahppens in either direction, or someone just commits and people bitch and moan Apr 28 15:41:22 yeah, it seems. I honestly have no idea whether the core team has ever actually made a single decision in its official capacity. Apr 28 15:42:18 um Apr 28 15:42:27 basically, we have an eV in place Apr 28 15:42:45 and there are some procedures in it for selecting members and working out voting procedures Apr 28 15:42:52 when is LinuxTAG and how many OE devs will be there? Apr 28 15:43:11 it seems like the next sensible oppurtunity to discuss properly Apr 28 15:43:24 http://www.linuxtag.org/2009/ Apr 28 15:43:39 we would need to do it b4, I am on vacation afterward Apr 28 15:44:05 XorA: starts 24- june-2009 Apr 28 15:44:39 * Crofton|work heads for lunch Apr 28 15:45:04 but Im for transferring control to eV members ASAP Apr 28 15:46:34 XorA: we have 4 eV members only Apr 28 15:47:17 hrw: I meant after we enrole eneryone who wants to join Apr 28 15:47:20 Florian, Holger, Mickeyl, Robert Apr 28 15:47:33 no Apr 28 15:47:35 we have more members Apr 28 15:47:36 huh, I am not a member anymore? Apr 28 15:47:40 these four are the voted board Apr 28 15:47:50 poor them :) Apr 28 15:47:51 everyone at the meeting is a member Apr 28 15:47:53 all are members who signed it Apr 28 15:47:57 correct Apr 28 15:48:13 I'm for some community selected group of steering groups, I think we are doing quite well in general btw. Apr 28 15:48:13 but its time to open up membership and transfer control Apr 28 15:48:24 what does "member" imply in this context? a stake in the decision making process? Apr 28 15:48:46 kergoth: no idea, my german is poor Apr 28 15:48:51 hehe Apr 28 15:48:58 *FG* Apr 28 15:49:12 we could have sold our souls for all the sense I could make of the Statutes Apr 28 15:49:15 kergoth: eV is the official body. it has nothing to do with commit policies and forwarding our metadata, luckily enough Apr 28 15:49:23 * kergoth nods Apr 28 15:49:31 XorA: well, then your German *is* good after all :-) Apr 28 15:49:41 thats why i asked, wondering exactly what "member" means in eV context Apr 28 15:49:51 XorA: you sold 4 of your lives, not all Apr 28 15:49:58 hehe Apr 28 15:50:08 hmm google OE eV -> http://www.tierschutzverein-oe.de/ *G* Apr 28 15:50:14 not that I mind losing my soul, but id prefer her to be hot :-) Apr 28 15:50:45 XorA: and she must come with her twin sister Apr 28 15:51:18 XorA: I sent the same patch for .dev Apr 28 15:51:28 http://en.wikipedia.org/wiki/Eingetragener_Verein Apr 28 15:52:20 http://de.wikipedia.org/wiki/Verein (de) Apr 28 15:52:22 XorA: how do things work then? Do I have to wait until the .dev patch is submitted? Apr 28 15:53:21 XorA: and then submit the stable again? Apr 28 15:53:47 member just means 'member of a voluntary association' in this context Apr 28 15:53:58 * mckoan wonders how OE patch policies really works :-) Apr 28 15:54:07 mckoan: commit and find out Apr 28 15:54:31 the mob will find you once you screwed up Apr 28 15:54:33 mckoan: yes Apr 28 15:55:50 * likewise_ reboots, hopes to have fixed sound under Jaunty finally. Apr 28 16:02:22 Tartarus: acked Apr 28 16:02:29 thanks Apr 28 16:02:47 that makes one :) Apr 28 16:11:34 bye for now Apr 28 16:12:23 mickeyl: any experience with telepathy for voip ? Apr 28 16:13:11 no. telepathy scares me Apr 28 16:13:13 :) Apr 28 16:14:37 so: experience yes: scared yes . (:-)) Apr 28 16:15:06 no, i looked at the docs Apr 28 16:15:09 and that alone scared me Apr 28 16:15:12 it's HUGE Apr 28 16:16:39 mickeyl: voip is ugly and complex Apr 28 16:19:59 re Apr 28 16:20:18 hi woglinde Apr 28 16:20:31 jo pb Apr 28 16:23:55 * pb_ go home now Apr 28 16:23:55 later all Apr 28 16:24:03 till later pb Apr 28 16:28:49 cu pb_ Apr 28 17:05:54 03Tom Rini  07org.openembedded.dev * r27d8943eab 10openembedded.git/contrib/buildbot/master.cfg: buildbot master.cfg: Fix location of Angstrom images for cleaning and full builds start from scratch Apr 28 17:16:31 Hi I compiled sdlquake with oe but I am just unable to run the application on my board :-( Apr 28 17:16:47 What happens? Apr 28 17:17:06 I am manually installing all the depencies but still there are some files missing Apr 28 17:17:25 and The ipk package is empty I think Apr 28 17:17:46 I did a bitbake sdlquake Apr 28 17:19:13 I copied the tmp/work/armv7a-angstrom-linux-gnueabi/sdlquake-1.0.9-r0/install/usr/bin/sdlquake image + the SDL lib Apr 28 17:19:38 NOw I have got this Apr 28 17:19:39 # sdlquake Apr 28 17:19:40 FindFile: can't find gfx/pop.lmp Apr 28 17:19:40 Playing shareware version. Apr 28 17:19:40 FindFile: can't find gfx.wad Apr 28 17:19:41 Error: W_LoadWadFile: couldn't load gfx.wad Apr 28 17:19:49 Well, you want to install the sqlquake ipk Apr 28 17:20:04 Yeah exactly Apr 28 17:20:23 So copy tmp/deploy/glibc/ipk/armv7a/sqlquake_... Apr 28 17:20:42 re Apr 28 17:20:57 hi florian Apr 28 17:21:05 Any chance you can review my SDK/stable patchset? :) Apr 28 17:21:29 or PREFERRED_LIBC rfc :) Apr 28 17:21:59 sorry not today, I just came back (!) to the office and have some hours of work to do :-( Apr 28 17:22:09 k Apr 28 17:22:47 I need to copy this sdlquake .ipk package on the target Apr 28 17:23:04 and then what is the command to install it ? Apr 28 17:23:13 ar -x ...ipk ? Apr 28 17:26:39 opkg install or so Apr 28 17:28:21 Thx Apr 28 17:28:33 you might have some luck in #angstrom, btw, that's where people using that distro hang out Apr 28 17:28:38 this is a bit more of a general OE dev channel Apr 28 17:41:11 re Apr 28 17:59:19 Ah yes, uclibc SDKs fail in stable until the patch i posted last night goes into dev then stable :( Apr 28 18:00:50 Tartarus: pthread debug one? Apr 28 18:00:55 yeah Apr 28 18:01:22 my new waterfall has red now :| Apr 28 18:02:40 tartarus which patch? Apr 28 18:02:48 the thread-db? Apr 28 18:02:59 yeah Apr 28 18:03:03 Tartarus: why the name PREFERRED_LIBC? Can the framework fall back to non-preferred automatically? Apr 28 18:03:12 Subject: [oe] [PATCH] Enable PTHREADS_DEBUG_SUPPORT on all uClibc builds Apr 28 18:03:13 one Apr 28 18:03:24 tartarus hm disabled it some times ago Apr 28 18:03:28 but there was a reason Apr 28 18:03:37 maybe it changed with version 0.9.30 Apr 28 18:03:38 likewise_, the goal is to make it easier to do uclibc overrides, and add in another libcs overrides if needed Apr 28 18:03:40 http://patchwork.openembedded.org/patch/425/ one Apr 28 18:04:11 tartarus the reason was if you enabled one debug option the whole build was build with debug flags Apr 28 18:04:43 woglinde, right. But we strip debug stuff out. Or are there runtime debug things too that we won't catch? Apr 28 18:05:21 likewise_, a distro could opt to set everything itself, but there's actually a number of silly mistakes being copy/pasted around Apr 28 18:05:25 tartarus optimazation flags was disbaled too Apr 28 18:05:42 likewise_, ie the ${TARGET_VENDOR} junk in preferred provider for ...libc-for-gcc Apr 28 18:06:12 tartarus I will check it in I think or didnt it khem already? Apr 28 18:06:20 woglinde, I only grepped around a recent one, I guess I need to fetch & unpack some older ones too Apr 28 18:06:32 woglinde, khem said on irc he'd look into it Apr 28 18:06:41 but I hit it again setting up buildbot for just SDK stuff Apr 28 18:06:51 hm Apr 28 18:08:10 Tartarus: I think the idea is great, just wondered about the "PREFERRED" part in the name. Apr 28 18:08:43 likewise_, well, I'm just stealing sane-toolchain.inc's name for it. I'm open to suggestions Apr 28 18:08:56 It's not a magic var like PREFERRED_PROVIDER/VERSION Apr 28 18:09:05 Tartarus: I mean, nobody is going to prefer uclibc, then finds it OK if the system overrides it to glibc. Apr 28 18:09:27 likewise_, heh Apr 28 18:09:47 It's just a flag for picking up the include file that sets PREFERRED_PROVIDER/foo right for that libc Apr 28 18:09:54 and adds libc to the OVERRIDES Apr 28 18:09:57 Tartarus: indeed, PREFERRED_ currently means: use this (version, provider) if available, fallback otherwise. We probably don't have nor want that behaviour for the C library Apr 28 18:10:35 likewise_, LIBC_PROVIDER, so it's not confused with the PREFERRED magic variables? Apr 28 18:10:41 I suck at naming Apr 28 18:11:18 me too, that's why I want them to be consistent :-) Apr 28 18:11:48 if I named OE vars, they would be WONE, TWO, THREE, FOUR ..... Apr 28 18:11:57 XorA|gone: Crofton|work I don't think we can be waiting for the next "in person" meeting to be moving on eV -- why can't we do everything with conf calls, etc? Apr 28 18:12:38 cbrake, we need to see how much of the kde articles we copied :) Apr 28 18:12:57 I do not know how it is set up to add members Apr 28 18:13:16 Crofton|work: we can have the board look at this. Apr 28 18:13:22 yeah Apr 28 18:13:38 does anyone on the board actually have time? Apr 28 18:13:54 I agree with cbrake, just move on with decision making. the board is mainly involved with legal matters / eV related. Apr 28 18:14:01 I'm hoping Robert pokes his head up Apr 28 18:14:12 so do I Apr 28 18:14:24 yeah me too, I just want to make sure we finish the eV stuff Apr 28 18:14:52 * cbrake sets up a git repo and sends Robert an email directly ... Apr 28 18:14:54 right now, we badly need the docs and an English translation :) Apr 28 18:15:17 heh Apr 28 18:15:21 good idea :) Apr 28 18:15:24 Crofton|work: what is it you want to know? Apr 28 18:15:42 how to add members :) Apr 28 18:15:58 Crofton|work: Do you have the eV docs? Apr 28 18:16:04 no Apr 28 18:17:26 woglinde, summarized what I found and posted to the ml Apr 28 18:18:16 hrw, is uclibc + vortex86sx valid? Apr 28 18:18:36 Tartarus: no, it is broken Apr 28 18:18:46 hrw, k, I'll remove from my list for now then Apr 28 18:24:47 bye Apr 28 18:39:54 hi, anyone here on Ubuntu 9.04 who has a solution for the dreaded "Failed to contact the GConf daemon; exiting." bug? Apr 28 18:40:13 (causes menuconfig and devshell tasks to fail) Apr 28 19:18:24 exit Apr 28 19:43:18 khem: failed to execute Apr 28 19:44:42 likewise_: ? Apr 28 19:44:51 I missed the context Apr 28 19:45:15 khem, you typed "exit" a while back Apr 28 19:45:37 heh Apr 28 19:45:38 hey khem Apr 28 19:45:53 I've got 2 acks, but can you ack the uclibc PTHREAD_DEBUG_SUPPORT change too please? Apr 28 19:46:02 I wanna get that into dev then push for stable to fix meta-toolchains Apr 28 19:46:06 likewise_: heh :) Apr 28 19:46:24 Tartarus: sure, I am lagging a bit Apr 28 19:51:01 khem: :-) Apr 28 19:52:21 khem: while you're at it, could think about tartarus' PREFERRED_LIBC implementation. I like the implementation, we just discussed that PREFERRED_ might imply it is a magic variable, which it is not, and we would like to hear your namespace expertise :-) Apr 28 20:25:04 Hmm Apr 28 20:25:28 When a stable branch maintainer proposes changes to what they maintain, still need 2 acks or just 1 ack from another stable maintainer? Apr 28 20:28:27 heh Apr 28 20:28:30 one ack Apr 28 20:28:34 unless core change :) Apr 28 20:28:59 * Tartarus goes back to thumb twiddling then Apr 28 20:29:07 although mckoan|away's kaelios specific stuff may suffer from lack of people familliar enough to ack Apr 28 20:29:36 Once we get beyond the generic distro? hard to say Apr 28 20:30:40 yeah Apr 28 20:30:56 I don't see this as a critical detail :) Apr 28 20:31:32 hm? Apr 28 20:31:40 any "commercial" business would filter even .stable in order to do quick fixes and make sure bad stuff did not creap in Apr 28 20:31:41 If a maintainer needs 2 acks or 1? Apr 28 20:31:42 or something else Apr 28 20:32:07 getting acks for pieces where not many people work Apr 28 20:32:15 am I making sense Apr 28 20:32:20 the "edges" Apr 28 20:32:25 Well, once we set a policy on it, sure Apr 28 20:32:29 yeah Apr 28 20:32:38 we will need to cover the case Apr 28 20:32:50 I'd be OK with 1 other maintainer acking when a maintainer proposes changes to what they own Apr 28 20:32:51 where we have trouble finding people to ack things Apr 28 20:32:59 but i'm biased, my SDK stuff might otherwise rot ;) Apr 28 20:33:05 yeah Apr 28 20:33:12 I do not want that to rot Apr 28 20:33:46 I wish I knew more about the sdk stuff Apr 28 20:34:11 once we have good doc on how to do sw dev for OE created systems Apr 28 20:34:11 It's not terribly different from normal toolchain stuff, once you get the hang of it :) Apr 28 20:34:18 things will get exciting ... Apr 28 20:34:31 I don't have the hang of that either :) Apr 28 21:01:30 cbrake: I hate phones :-D Apr 28 21:01:55 cbrake: but no actual issue, it just seems to me OE devs get on better in the presence of beer Apr 28 21:02:34 XorA|gone: heh, yeah I can imagine -- sounds like a few would help now!! Apr 28 21:02:48 heh Apr 28 21:03:08 generally, you distract them with shiny gadgets Apr 28 21:03:45 I forget what we actually put in the articles of association about meetings Apr 28 21:09:51 03Tom Rini  07org.openembedded.dev * r7e110765a9 10openembedded.git/recipes/uclibc/ (17 files in 11 dirs): (log message trimmed) Apr 28 21:09:51 uClibc: Always enable PTHREADS_DEBUG_SUPPORT Apr 28 21:09:51 This will make sure we always have the uclibc-thread-db package, which Apr 28 21:09:51 task-sdk-bare requires. After an audit of the versions in question, what Apr 28 21:09:51 always happens now is that all libpthread debugging stuff from libpthread Apr 28 21:09:55 ends up in a specific file and package. Apr 28 21:09:57 Acked-by: Koen Kooi Apr 28 21:36:45 * Jay7 -> sleep Apr 28 22:02:30 kergoth: ping Apr 28 22:02:47 again sh: rm: command not found Apr 28 22:03:11 it happens during task do_setscene Apr 28 22:03:17 again, why are you telling me this? Apr 28 22:03:20 i didn't write that code Apr 28 22:03:27 'cause seems bitbake problem... Apr 28 22:03:30 no, it isnt Apr 28 22:03:33 base.bbclass? Apr 28 22:03:34 its an openembedded problem Apr 28 22:03:36 http://fr.pastebin.ca/1406686 Apr 28 22:03:49 base.bbclass isnt bitbake, its a core class in openembedded Apr 28 22:03:59 do_setscene was done by rp, i think, when he added packaged staging Apr 28 22:03:59 ok, sorry... Apr 28 22:04:19 you are the only one understanding well both :-) Apr 28 22:05:44 it seems oe_libinstall pewrhaps... Apr 28 22:08:44 this perhaps ./run.staging_helper.4037: echo "Removing $dir" ; rm $dir -rf Apr 28 22:09:02 rm -rf $dir seems better... Apr 28 22:10:11 'rm -rf $dir' is certainly more compliant with the letter of the law, but I would be surprised if the other form caused 'rm: command not found' errors. Apr 28 22:10:20 sounds like that must be caused by a broken $PATH or some such thing Apr 28 22:10:55 hm.../run.do_rm_work.4203: __runcmd rm -f $destpath/$libname.la Apr 28 22:10:55 anyway, try running that script with 'set -x' and see what happens. Apr 28 22:11:31 pb_: the error is sh: rm: ... Apr 28 22:11:34 so is shell Apr 28 22:11:46 right, but the whole thing is a shell script Apr 28 22:11:51 perhaps __runcmd Apr 28 22:11:52 all the run.do_xx scripts are shell scripts Apr 28 22:12:01 I'll paste the runlogs Apr 28 22:12:12 I'm just grepping there :-) Apr 28 22:12:47 a simple thing like bitbake package-index should show the issue Apr 28 22:12:57 there too Apr 28 22:13:02 there probably isn't much point pasting them, I doubt it will be easy to spot the problem by inspecting the files. you really need to run the thing and find out where it's failing. Apr 28 22:13:30 as I say, run it under -x (or, even better, -x -e) and that should reveal pretty quickly which line is going wrong. Apr 28 22:39:00 Is there a nice way to reconfig a kernel with menuconfig after it's been built once? Apr 28 22:41:01 bitbake -c menuconfig ? Apr 28 22:42:15 will the "-c" clean out the kernel source and re-untar/patch it or just reconfig in place? Apr 28 22:42:37 no Apr 28 22:42:51 '-c' is prefix for 'command' Apr 28 22:43:01 ah Apr 28 22:43:23 bitbake -c clean virtual/kernel doeas it Apr 28 22:44:43 I see. The "-c menuconfig" gave me "Nothing to do" (using arago btw) Apr 28 22:47:31 for a kernel, there is a task menuconfig Apr 28 22:47:32 never mind I got it Apr 28 22:47:38 Thanks Apr 28 22:47:41 np Apr 28 22:48:57 03Tom Rini  07org.openembedded.dev * r07e54135f7 10openembedded.git/classes/insane.bbclass: (log message trimmed) Apr 28 22:48:57 insane.bbclass: Split package_qa_check_devdbg and don't run -dev check on SDK packages. Apr 28 22:48:57 We create package_qa_check_dev/package_qa_check_dbg checks. While in here Apr 28 22:48:57 clarify the comments as to what both are doing. Since SDK packages don't Apr 28 22:48:57 get split like normal ones, it's OK to have -dev stuff in non-dev packages. Apr 28 22:49:01 Suggested by Richard Purdie Apr 28 22:49:03 Acked-by: Koen Kooi Apr 28 22:57:21 03Andrea Adami  07org.openembedded.dev * r6422fe2b0a 10openembedded.git/recipes/kexecboot/linux-kexecboot.inc: Apr 28 22:57:21 linux-kexecboot.inc: force CONFIG_KEXEC=y just in case Apr 28 22:57:21 - no runtime changes, no PR BUMP Apr 28 22:59:09 hm. After menuconfig finishes, how to build and deploy? Apr 28 23:03:42 hoj: 1) copy the .config out of workdir and create a patch 2)bitbake -f -c compile virtual/kernel Apr 28 23:04:07 remember you'll loose your changes in workdir! Apr 28 23:05:12 ant__: Hm, thanks. So no way to recompile in-place without reextracting+patching? Apr 28 23:05:23 yes, 2) Apr 28 23:05:32 -f = --force Apr 28 23:05:48 you restart from the specific task Apr 28 23:06:09 try bitbake -c listtasks virtual/kernel Apr 28 23:06:48 read bitbake man :-) Apr 28 23:06:55 I have to go, is late here... Apr 28 23:07:00 good night Apr 28 23:49:17 Crofton|work: who is on the core team? Apr 28 23:57:48 rwhitby, that is a good question :) Apr 28 23:57:56 people on oe-private I guess Apr 28 23:58:12 I'm in the middle of some other stuff Apr 28 23:58:34 then you have the subset that posts ... Apr 28 23:59:06 and then you have the subset of people willing to help with solutions and supply money :) Apr 29 00:10:51 Crofton: OE is a big community - money should not be a problem. nslu2-linux has spent over USD$10K of community donations on server hardware in the last 3 years. Apr 29 00:16:17 Is Philip Balister on IRC? Apr 29 00:28:00 rwhitby, yes :0 Apr 29 00:29:00 but, I am installing mailbox ,,,,, Apr 29 00:30:48 er snail mail box Apr 29 00:31:32 either can be a bitch ;) Apr 29 00:36:12 Crofton: oh, sorry - I hadn't made the nick<->realname match mentally. Apr 29 01:01:13 03Jeremy Lainé  07fso/milestone5.5 * r8091ef79b2 10openembedded.git/recipes/gnome/libsoup-2.4_2.26.0.bb: libsoup-2.4: add missing DEPENDS on sqlite3 Apr 29 01:01:13 03Koen Kooi  07fso/milestone5.5 * r90d550951a 10openembedded.git/recipes/ (4 files in 3 dirs): libsoup: fix packaging to keep upgrade paths sane, bump PR on affected packages Apr 29 01:01:15 03Koen Kooi  07fso/milestone5.5 * r17499bcb27 10openembedded.git/recipes/gnome/libsoup-2.4_2.26.0.bb: libsoup: add 2.26.0 Apr 29 01:01:15 03Koen Kooi  07fso/milestone5.5 * re53759d2d1 10openembedded.git/ (conf/checksums.ini recipes/gnome/libproxy_0.2.3.bb): libproxy: add 0.2.3 Apr 29 01:01:18 03Koen Kooi  07fso/milestone5.5 * r5b4f4c8d9b 10openembedded.git/ (6 files in 3 dirs): Apr 29 01:01:20 webkit-gtk: switch to libsoup backend (segfaults are gone) to get cookies working with midori Apr 29 01:01:22 * runtime tested on beagleboard Apr 29 01:01:24 03Koen Kooi  07fso/milestone5.5 * rdd7f01fc34 10openembedded.git/recipes/gnome/libproxy_0.2.3.bb: libproxy: add libxmu to depends Apr 29 01:01:29 03Angus Ainslie  07fso/milestone5.5 * rf17e472098 10openembedded.git/conf/distro/include/sane-srcrevs.inc: sane-srcrevs : bump frameworkd Apr 29 01:01:32 03Angus Ainslie  07fso/milestone5.5 * r7adc776a6a 10openembedded.git/recipes/e17/eve_svn.bb: eve: fix SRC_URI and inherit e.bbclass Apr 29 01:01:35 03Angus Ainslie  07fso/milestone5.5 * r01170ae550 10openembedded.git/recipes/webkit/webkit-efl_git.bb: webkit-efl : add libsoup-2.4 and inherit pgktools Apr 29 01:01:38 03Angus Ainslie  07fso/milestone5.5 * r8987b98249 10openembedded.git/recipes/connman/ (connman.inc connman_0.15.bb): connman : add dbus-services file Apr 29 01:06:04 rwhitby: how has your experience been with hosting at OSUOSL? Apr 29 01:06:23 cbrake: perfect Apr 29 01:09:52 no unplanned outages, plenty of notice of planned possible outages (none of which took our sites offline) and they've been very helpful with hardware upgrades Apr 29 01:19:41 rwhitby: excellent -- do you have at least two contacts that are interested in helping you resolve issues? Apr 29 01:20:32 cbrake: there is a hosting mailing list and defined support ticket system with multiple admins resolving issues. Apr 29 01:20:49 we've actually never had an unplanned issue that they had to resolve for us. only hardware and disk upgrades. Apr 29 01:39:04 Crofton: I have also contacted softwarefreedom.org for advice on non-profit issues Apr 29 02:32:52 cbrake_away, that seems familiar **** ENDING LOGGING AT Wed Apr 29 02:59:57 2009