**** BEGIN LOGGING AT Wed Apr 27 02:59:58 2011 Apr 27 04:25:00 Where can I find standard procedures for contributing to OE? Apr 27 04:25:20 JavaRockyMobile: look at wiki Apr 27 04:25:51 Commit policy? Apr 27 04:26:22 I see, exactly what I am looking for. Cheers. Apr 27 04:27:39 http://openembedded.org/index.php/New_Dev Apr 27 04:28:04 there are many helpful docs in wiki Apr 27 04:28:58 Oh yes, thanks. I havent ventured into the dev area as yet. Will do now! Apr 27 07:10:52 I want to know which package prints "shared library dependencies " ..... as in ubuntu it is provided by 'ldd' Apr 27 07:11:21 I had builded a filesystem using oe for mini2440... Apr 27 07:18:41 ldd package is built as part of target binutils Apr 27 07:18:44 sanket: ^ Apr 27 07:31:13 hej, anybody using bblayers noticed that there is something wrong with distutils based recipes, in the sense that if(f) you build third party module it brings in .L/usr/lib instread of the staging lib/ ? (I think it searches the wrong python-config, but I am not sure) Apr 27 07:32:08 yes Apr 27 07:33:02 mlip: http://www.mail-archive.com/openembedded-core@lists.openembedded.org/msg00201.html Apr 27 07:35:00 JaMa|Wrk, I did tinker in the recipes and distutils package yesterday, and I think this doesn't completelly solve it? distutils appends a library dir itself when creating the linking command Apr 27 07:36:11 JaMa|Wrk, https://bitbucket.org/carljm/python-distutils/src/48c42eeaee44/command/build_ext.py#cl-286 Apr 27 07:36:54 mlip: this is more like workaround I used to push sysroot libdri before any other libdir Apr 27 07:37:27 but it was enough for couple of 3rdparty modules and that's what we have in oe.dev Apr 27 07:37:36 JaMa|Wrk, ah, ok; the problem is somewhere in distutils, because it seems to grab (or even mix, idk) the Makefile configurations Apr 27 07:39:23 maybe I should write to the ML and someone with more distutils experience can fix it looking at the right Makefile Apr 27 07:56:31 JaMa|Wrk, maybe you can help me out here: the culprit seems to be that STAGING_LIBDIR points into sysroots/-/usr/lib Apr 27 07:56:45 while it should somehow point into some cross dir Apr 27 07:56:59 is there some STAGING_LIBDIR_CROSS ? Apr 27 08:22:36 JaMa|Wrk, ping? Apr 27 08:26:04 mlip: sorry bussy daywork Apr 27 08:28:09 JaMa|Wrk, np; if you have a minute: the culprit is that STAGING_LIBDIR points into target lib (thats correct), but it then parses python/config/Makefile which is the one that is installed on the target containing /usr/lib as libdir; we could patch this file in some do_stage_append ? (as we don't want to crush anything on the target itself) Apr 27 08:29:08 (or some kind of sysroot append) Apr 27 08:31:17 you can mangle it from SYSROOT_PREPROCESS_FUNCS like binconfig.bbclass does, but maybe it's better to prefix value parsed from python/config/Makefile with sysroot Apr 27 08:37:12 morning all Apr 27 08:41:13 is there an integrated mechanism to strip build libs/tools in an own recipe? Apr 27 08:44:01 iirc, libs and binaries are stripped automatically Apr 27 08:44:26 debug symbols going to -dbg package Apr 27 09:13:28 /join #angstrom Apr 27 09:22:32 03Martin Jansa  07master * rd68a33bedd 10openembedded.git/recipes/bluez/bluez4_4.91.bb: Apr 27 09:22:32 bluez4-4.91: bump PR to rebuild after e557584a06b312e19811e07fdf75d1a7f123ed91 Apr 27 09:22:32 Signed-off-by: Martin Jansa Apr 27 11:28:35 khem ping Apr 27 11:59:52 hello. Apr 27 12:00:13 is it possible to set SRCREV to a branch name in _git bitbake recipes? Apr 27 12:01:13 CMoH-office: you should set branch in SRC_URI and rev only in SRCREV Apr 27 12:10:53 ty JaMa|Wrk, what i'm aiming for is for the git recipe to retrieve the head of a branch. however, not providing any SRCREV, using SRCREV="HEAD" or using SRCREV="" all fail Apr 27 12:12:51 CMoH-office: SRCREV=AUTOREV Apr 27 12:13:56 CMoH-office: put known to work revision to in recipe SRCREV and add SRCREV_pn-your-blah = "${AUTOREV}" to some .inc file or your distribtuion config Apr 27 12:13:58 hmm... last advice i got was SRCREV="${AUTOREV}", which works for SVN but not git. are you referring to a SRCREV=AUTOREV without the ${} variable expansion syntax? Apr 27 12:14:18 it works for svn as well as for git Apr 27 12:14:36 i see; i'll try again and see if it wasn't some other problem at the time i've tested Apr 27 12:14:38 and with ${} Apr 27 12:14:48 thanks JaMa|Wrk :) Apr 27 14:02:17 hi , is openembedded supporting leopardboard dm368 ? Apr 27 14:03:32 hi , is openembedded supporting leopardboard dm368 ? Apr 27 14:06:16 hi , is openembedded supporting leopardboard dm368 ? Apr 27 14:09:14 *sigh* Apr 27 14:17:16 jama it works this way I hope Apr 27 14:17:21 hi , is openembedded supporting leopardboard dm368 ? Apr 27 14:20:07 change nick ! Apr 27 14:21:15 or change channel :) Apr 27 14:21:58 ah right I forget it Apr 27 14:31:31 what Apr 27 14:31:41 woglinde , leopardboard 355 ? Apr 27 14:33:28 hi , is openembedded supporting leopardboard dm368 ? Apr 27 14:41:15 woglinde , u ask same question ? Apr 27 14:41:18 whoh Apr 27 14:41:31 MACHINE=dm355-leopard Apr 27 14:41:42 dm355 and dm368 same ? just speed difference ? Apr 27 14:44:29 hi , is openembedded supporting leopardboard dm368 ? Apr 27 14:51:36 where id duke nukem when we need him :) Apr 27 14:51:39 is Apr 27 14:52:30 03Andrea Adami  07org.openembedded.dev * r0fad563e9a 10openembedded.git/recipes/linux/linux_git.bb: Apr 27 14:52:30 linux_git: move to 2.6.39-rc5 Apr 27 14:52:30 Signed-off-by: Andrea Adami Apr 27 14:52:38 03Andrea Adami  07org.openembedded.dev * rcadda2b6c4 10openembedded.git/recipes/linux/linux-kexecboot_git.bb: Apr 27 14:52:38 linux-kexecboot_git: move to 2.6.39-rc5 Apr 27 14:52:38 Signed-off-by: Andrea Adami Apr 27 14:53:00 woglinde: pls check if simpad boots ^^^ Apr 27 14:57:44 lhi , is openembedded supporting leopardboard dm368 ? Apr 27 14:57:53 * tdebrouw just had to ask it... Apr 27 14:58:40 maybe we should ask this everyday between 16h30 - 17h :) Apr 27 14:58:50 as an internal joke Apr 27 14:59:52 *between 16h -> 17h Apr 27 15:01:27 * tdebrouw just received a guruplug display Apr 27 15:01:37 now you be jealous :P Apr 27 15:01:40 ant? Apr 27 15:01:46 ah you mean for .39 Apr 27 15:24:54 hi , is openembedded supporting leopardboard dm368 ? Apr 27 15:25:31 ant__: seems you are have some free time :) Apr 27 15:26:16 only this week ;) Apr 27 15:27:22 some kind of vacation? Apr 27 15:27:54 maybe ant__ has taken time off for the royal wedding ;) Apr 27 15:28:03 wife + kid are away one week ;) Apr 27 15:28:13 ant__: ah, I understand ;) Apr 27 15:28:21 I'm doing some recabling at home Apr 27 15:28:39 better to be alone... Apr 27 15:28:46 exposed 220AC Apr 27 15:29:04 floating terminale... Apr 27 15:29:21 tools everywhere? Apr 27 15:30:17 heh... Apr 27 15:38:35 bluelightning: nice purge! Apr 27 15:39:11 ant__: thanks... there are a few bits yet to be removed but that zaps most of it Apr 27 15:39:47 would you please run the scripts against oe-dev? I'm trying to separe good stuff to keep from old cruft... Apr 27 15:40:12 ant__: I didn't have any scripts for this one, it was all done by hand :o Apr 27 15:40:33 he, I feared Apr 27 15:40:51 grep & co. Apr 27 15:40:56 yep Apr 27 15:41:11 it's also a lot easier for oe-core as being based on poky the number of distros/machines originally in it were a small subset of OE's Apr 27 15:41:19 there is some 2.4 stuff in oe-dev tasks and recipes Apr 27 15:41:53 yes, I think there's still 2.4 stuff in oe-core... is 2.4 completely deprecated in oe-dev? Apr 27 15:41:53 bluelightning, have you got a source push for 125 yet? Apr 27 15:41:57 I don't know if it's worth cleaning it in oe-dev or just wait and port only the sane bits to meta-oe Apr 27 15:42:37 bluelightning: iirc 2.4 is unsupported since one of first TSC meetings Apr 27 15:42:49 ka6sox-away: I thought we already sorted that out... ? Apr 27 15:43:12 I thought we had 124 but I didn't see one for 125 Apr 27 15:43:14 there were actually onle a few machines still on linux24 Apr 27 15:43:49 whose stuff was moved by me to nonworking/obsolete Apr 27 15:44:09 now we have a lot of stale overrides Apr 27 15:46:17 ant__: ah I see Apr 27 15:46:51 ant__: the easiest approach would be to just port the sane bits I guess Apr 27 15:47:08 although thinning out the old cruft would make it easier to do that Apr 27 15:50:36 ka6sox-home: unless I'm misunderstanding, the 1.2.5 tarballs are already up on sources.openembedded.org... ? Apr 27 15:53:02 bluelightning: okay...I'll poke around and see if its all there..thanks Apr 27 15:53:26 ka6sox-home: ok give me a shout if anything seems amiss Apr 27 15:58:32 I need to add a custom fstab for my machine, but want to do it in the machine overlay, rather than add it to base files Apr 27 15:58:43 what's the best mechanism for that Apr 27 15:58:55 base-file Apr 27 15:59:43 http://sakrah.homelinux.org/blog/2010/11/bblayers-bbappend/ Apr 27 16:31:41 khem, how did the timing go on the last run? Apr 27 16:32:09 looking ... Apr 27 16:32:17 khem, 1 arch used 41% of disk. Apr 27 16:32:22 ok Apr 27 16:32:26 329m17.036s Apr 27 16:32:39 so not faster... Apr 27 16:32:53 yeah about 5:30 hrs Apr 27 16:33:01 since you did 5 in 1300 minutes Apr 27 16:33:23 well subsequent build sshare the native sysroot Apr 27 16:33:29 so they will be faster Apr 27 16:33:32 kk. Apr 27 16:33:39 e.g. on top of this build if I build for mips Apr 27 16:33:45 it will be lot faster now Apr 27 16:33:53 since it does not ahve to do common parts Apr 27 16:34:15 ok 41% is like 35G Apr 27 16:34:16 so we can do 2@ a time. and still fit....maybe 3 Apr 27 16:34:20 yes Apr 27 16:34:33 ok incrementally let me do one for mips Apr 27 16:34:38 and see where we stand Apr 27 16:34:39 kk Apr 27 16:34:50 then if we still have space then I can do for ppc Apr 27 16:34:53 and so on Apr 27 16:34:58 right Apr 27 16:35:04 I think first build will eat the most Apr 27 16:35:17 since it builds the common parts too Apr 27 16:35:18 all the common bits. Apr 27 16:35:23 yep Apr 27 16:35:39 but I am disappointed that it takes 5:30 hrs Apr 27 16:35:47 for a console-image Apr 27 16:36:16 there is a lot still being done outside the work dir... Apr 27 16:36:21 but may be that will fuel some improvements to bitbake Apr 27 16:36:22 thats whats taking the time. Apr 27 16:36:36 yeah everything runs under pseudo Apr 27 16:36:47 which is kind of needed slowness Apr 27 16:36:59 you can see it in the HD logs. Apr 27 16:37:19 I launched mips build Apr 27 16:37:23 lets see how it does Apr 27 16:37:25 http://munin.nas-admin.org//openembedded.org/garnet/diskstats_utilization/index.html Apr 27 16:37:28 did you clean first? Apr 27 16:37:32 no Apr 27 16:37:35 no cleaning Apr 27 16:37:36 or just let it build Apr 27 16:37:47 I building on top Apr 27 16:38:00 that will tell us exaclty how much more disk it will take Apr 27 16:38:01 that previous was mips console image? Apr 27 16:38:05 if I clean then it wont Apr 27 16:38:11 since it will rebuild common parts Apr 27 16:38:14 kk Apr 27 16:38:19 khem: wrt your oe-core e-mail: shouldn't it be PREFERRED_PROVIDER-gconf-utils? or better virtual/gconf-utils? Apr 27 16:38:55 khem: as I understand it newer gconf provides gconf-utils, but gconf-utils does not provide full gconf functionality Apr 27 16:39:00 JaMa|Off: I think oe-core should just drop gconf-dbus Apr 27 16:39:31 well I have it blacklisted too... Apr 27 16:39:44 the reason I send it is so yocto folks can test it out Apr 27 16:39:57 since I dont test it in the same env as theirs Apr 27 16:39:58 and had to change it only in libproxy (because the rest is not in my dep. tree) Apr 27 16:40:18 if you build meta-toolchain-sdk Apr 27 16:40:21 then you will see Apr 27 16:40:29 I think there might be more references Apr 27 16:40:33 but I did not check Apr 27 16:41:14 ok, I don't really care.. just potentional PREFERRED_PROVIDER-gconf = gconf-utils looks wrong to me Apr 27 16:42:36 I have removed all refs Apr 27 16:42:49 poky.conf:PREFERRED_PROVIDER_gconf ?= "gconf-dbus" Apr 27 16:43:53 yes that's what looks wrong to me Apr 27 16:44:21 as recipes which had full gconf in DEPENDS before will get now gconf-dbus only Apr 27 16:45:16 if you replace every gconf-dbus with virtual/gconf-dbus and set PREFERRED_PROVIDER-gconf-dbus = "gconf" for angstrom, then everything should work right Apr 27 16:45:38 arg PREFERRED_PROVIDER-virtual/gconf-utils = "gconf" Apr 27 17:06:31 anyone have a good bbappend example Apr 27 17:06:50 FILESEXTRAPATHS := "${THISDIR}/${PN}" Apr 27 17:06:50 refereneced in the yocto bsp docs does not work in oe :( Apr 27 17:25:42 khem, your blog posting on bbappen is out of date Apr 27 17:25:51 I get syntax errros with it Apr 27 17:44:41 Crofton|work: thanks, I should update it Apr 27 17:47:55 khem, any suggestion would it should be to overwrite a conf file from an overlay? Apr 27 17:50:21 the example in the yocto bsp does not work, which is unfortunate ... Apr 27 17:53:10 03Denis 'GNUtoo' Carikli  07master * rceb2a05d0a 10openembedded.git/conf/machine/nokia900.conf: Apr 27 17:53:10 nokia900: change to newer 2.6.37 kernel which supports power management Apr 27 17:53:10 Signed-off-by: Denis 'GNUtoo' Carikli Apr 27 17:53:35 03Denis 'GNUtoo' Carikli  07master * r3e8b3e9ca7 10openembedded.git/recipes/linux/ (linux-nokia900-meego/defconfig linux-nokia900-meego_git.bb): Apr 27 17:53:35 add new linux-nokia900-meego recipe Apr 27 17:53:35 This kernel version(which is for the nokia900 machine) Apr 27 17:53:35 has power management(CPUFREQ,DVFS),along with support Apr 27 17:53:35 for the nokia900 and its drivers. Apr 27 17:53:36 Signed-off-by: Denis 'GNUtoo' Carikli Apr 27 17:53:39 03Klaus Kurzmann  07master * rbfd4e4d292 10openembedded.git/conf/machine/nokia900.conf: Apr 27 17:53:39 nokia900.conf: autoload isp1704_charger and omap_ssi modules Apr 27 17:53:39 Signed-off-by: Klaus Kurzmann Apr 27 17:53:55 03Denis 'GNUtoo' Carikli  07master * rc012be8a84 10openembedded.git/recipes/xorg-xserver/ (2 files in 2 dirs): Apr 27 17:53:55 xserver-xorg-conf: add support for newer kernel for the nokia900 Apr 27 17:53:55 Note that there are whitespaces issues in the nokia900's xorg.conf Apr 27 17:53:55 Also note that it may not be the final xorg.conf since the kernel Apr 27 17:53:55 defconfig may change to add new input devices(sensors etc...) Apr 27 17:53:55 Signed-off-by: Denis 'GNUtoo' Carikli Apr 27 17:59:08 Crofton|work: if you just want to use a different conf file, you can simply adjust your FILESPATHBASE to look in your overlay first Apr 27 17:59:29 it is not clear to me how to do that :) Apr 27 17:59:47 one sec, dialing into meeting Apr 27 18:02:39 Crofton|work: http://pastebin.com/PHudpGai Apr 27 18:02:58 I'm no expert (I found that stuff on IRC logs a while back), but it's working here. Apr 27 18:03:07 I put that in build/conf/local.conf Apr 27 18:03:10 what file did you put that in? Apr 27 18:03:15 ^ Apr 27 18:03:23 herh Apr 27 18:03:33 so COLLECTIONS has my overlay and OE Apr 27 18:04:04 it's voodoo and I never tried to fully understand it, but it "works for me"tm :) Apr 27 18:09:22 foerster, there needs to be a saner way :) Apr 27 18:09:35 Crofton|work: perhaps :) Apr 27 18:09:49 well, bbappend is the other approach Apr 27 19:37:39 03Lukas Märdian  07master * r65b9bb72df 10openembedded.git/recipes/shr/e-wm-config-illume2-shr_git.bb: Apr 27 19:37:39 e-wm-config-illume2-shr: bump SRCREV and PR Apr 27 19:37:39 * this replaces e's default battery gadget with the newly introduced battery-fso gadget (from shr-e-gadgets) Apr 27 19:37:39 * affects all shr powered devices Apr 27 19:37:39 * changes will be visible on the device after upgrade, once ~/.e has been removed Apr 27 19:37:39 Signed-off-by: Lukas Märdian Apr 27 19:37:47 03Lukas Märdian  07master * r829e4a5d3e 10openembedded.git/recipes/shr/shr-e-gadgets_git.bb: Apr 27 19:37:47 shr-e-gadgets: bump SRCREV and PR Apr 27 19:37:47 * this adds a new e-gadget: battery-fso Apr 27 19:37:47 * battery-fso is an alternative battery gadget, which uses fsodeviced as data source Apr 27 19:37:47 Signed-off-by: Lukas Märdian Apr 27 19:43:32 is there a recommended version of bitbake to go with the 2011.03 release of oe master? Apr 27 19:52:59 1.10-1.12 Apr 27 19:53:34 Jay7: thx Apr 27 20:32:37 re Apr 27 20:32:37 khem ping Apr 27 20:41:37 woglinde: yes Apr 27 20:48:14 khem could you take a look at uclibc backtrace for arm and eabi Apr 27 20:48:56 when compiling libbacktrace a eabi symbol is missing Apr 27 20:53:15 sure Apr 27 20:53:20 what error do u runinto Apr 27 20:53:26 is it happening with oe ? Apr 27 20:53:35 yes oe Apr 27 20:53:43 ok Apr 27 20:53:52 try with oe-core :) Apr 27 20:53:56 and meta-oe soon Apr 27 20:54:38 khem: what's maybe disturbing is the lack of a clear 'minimal' OE-branded distro-layer Apr 27 20:55:08 we are forced to use A. or shr if we want to play with layers today Apr 27 20:55:17 (if I'm not too wrong :) Apr 27 20:55:41 ant__: you could use oe-core only too Apr 27 20:56:07 and where to put OE logos for kernel, psplash, icons... Apr 27 20:56:09 ? Apr 27 20:56:20 oe-core probably Apr 27 20:56:31 is best place Apr 27 20:56:38 would be great Apr 27 20:56:57 just post patches for all this to oe-core Apr 27 20:57:29 ok, I'm summing up some Apr 27 20:57:37 ant__: good Apr 27 20:57:42 btw I still have one for oe-dev Apr 27 20:57:56 oerunmake savedefconfig Apr 27 20:58:06 I'll post this to oe-dev ML Apr 27 21:16:00 03Andrea Adami  07org.openembedded.dev * r77bb427ec8 10openembedded.git/recipes/linux/linux-2.6.38/ (6 files in 6 dirs): Apr 27 21:16:00 linux: update Zaurus 2.6.38 defconfigs Apr 27 21:16:00 * shrink using do_savedefconfig Apr 27 21:16:00 Signed-off-by: Andrea Adami Apr 27 21:31:24 Hi all , when i build my recipe it does not run more than one job . But with other recipe I am able to build multiple jobs . i also specified PARALLEL_MAKE = "-j 8" BB_NUMBER_THREADS = "8" in my local.conf file Here is my recipe http://pastebin.com/1ejGnFaB Apr 27 21:36:46 can any one point why i am not able to execute multiple jobs Apr 27 21:38:07 sreddy: if all other tasks already done only your recipe will be built Apr 27 21:38:54 yes but my recipe should have multiple jobs Apr 27 21:39:44 I'm not sure but inherit qt4e may set PARALLEL_MAKE to "" Apr 27 21:39:57 I remember some problems wrt qt and P_M Apr 27 21:40:11 Jay7: hmmm .. i will check that Apr 27 21:40:11 but this is just idea Apr 27 21:40:25 got it thanks Apr 27 21:51:52 03Andrea Adami  07org.openembedded.dev * r93db86243c 10openembedded.git/recipes/linux/linux/ (6 files in 6 dirs): Apr 27 21:51:52 linux_git: add Zaurus defconfigs for 2.6.39-rc5 Apr 27 21:51:52 * shrink using do_savedefconfig Apr 27 21:51:52 Signed-off-by: Andrea Adami Apr 27 22:07:27 03Andrea Adami  07org.openembedded.dev * r71eb55a998 10openembedded.git/recipes/linux/files/ (4 files in 2 dirs): Apr 27 22:07:27 linux: move stale patches to /obsolete/linux Apr 27 22:07:27 Signed-off-by: Andrea Adami Apr 27 22:07:29 03Andrea Adami  07org.openembedded.dev * rd13ba4be00 10openembedded.git/recipes/linux/linux/ (defconfig patch-2.6.28-gc): Apr 27 22:07:29 linux: move gamecube specific files to their dir Apr 27 22:07:29 Signed-off-by: Andrea Adami Apr 27 22:37:11 Hi all in qmake_base.bbclass EXTRA_OEMAKE = ' MAKEFLAGS= ' is overriding the make flags , and not able to build with multiple jobs after after setting PARALLEL_MAKE = "-j 8" BB_NUMBER_THREADS = "8" . has this issue been fixed in recent versions ? Apr 27 23:38:43 khem, 42.82% Apr 27 23:39:47 ka6sox-work: ok Apr 27 23:39:54 ka6sox-work: 259m27.530s Apr 27 23:40:11 so a little over 4hrs Apr 27 23:40:32 yes thats pretty much how much it will cost for rest of three Apr 27 23:40:54 4*4 + 6.5 Apr 27 23:41:05 makes is 22.5 hrs Apr 27 23:41:07 which is what we saw before 22.5hrs Apr 27 23:41:12 yes Apr 27 23:41:21 so in our case rm_work doesn't hurt us on time. Apr 27 23:41:34 so now we stand at 42.8% ? Apr 27 23:41:36 thats it Apr 27 23:41:41 yes Apr 27 23:42:07 so I suspect we can build 2 more and see if we have enough room for #5 Apr 27 23:42:13 we should. Apr 27 23:44:39 ka6sox-work: yes I have started ppc Apr 27 23:44:45 and will check back in 4 hrs Apr 27 23:45:09 okay, then we can do x86 and hopefully x86_64 isn't any bigger **** ENDING LOGGING AT Thu Apr 28 02:59:57 2011