**** BEGIN LOGGING AT Tue Jan 06 02:59:56 2009 Jan 06 07:58:05 morning all Jan 06 08:36:50 03Koen Kooi  07org.openembedded.dev * r5437295c58 10openembedded.git/packages/powervr-drivers/ (libgles-omap3/rc.pvr libgles-omap3_1.0.bb): libgles omap3: use insmod in initscript instead of modprobe to avoid segfault Jan 06 08:36:50 03Koen Kooi  07org.openembedded.dev * re41d119807 10openembedded.git/packages/linux/ (linux-omap-2.6.27/beagleboard/defconfig linux-omap_2.6.27.bb): linux-omap 2.6.27: enable bc_example for beagle and turn on sgx debugging Jan 06 08:36:51 03Koen Kooi  07org.openembedded.dev * r15e332a21b 10openembedded.git/packages/powervr-drivers/libgles-omap3_1.0.bb: libgles-omap3: work around SDK bugs in packaging Jan 06 08:36:54 03Koen Kooi  07org.openembedded.dev * r5ce58de452 10openembedded.git/packages/clutter/ (clutter_0.8+svn.bb clutter_svn.bb): clutter: add recipe to build from the 0.8 release branch Jan 06 08:36:57 03Koen Kooi  07org.openembedded.dev * rf946fc4c27 10openembedded.git/packages/clutter/clutter-box2d_svn.bb: clutter-box2d: bump PR for clutter change Jan 06 08:37:00 03Koen Kooi  07org.openembedded.dev * r65ee3358e6 10openembedded.git/packages/linux/ (linux-omap-2.6.27/beagleboard/defconfig linux-omap_2.6.27.bb): linux-omap 2.6.27: increase DMA size for beagleboard Jan 06 08:37:05 03Koen Kooi  07org.openembedded.dev * r299f1c2aba 10openembedded.git/packages/dsplink/ (codec-engine_2.21.bb gstreamer-ti_svn.bb ti-paths.inc): gstreamer-ti: add skeleton for TI gstreamer plugins Jan 06 08:37:10 03Michael 'Mickey' Lauer  07org.openembedded.dev * r004d985bef 10openembedded.git/ (2 files in 2 dirs): python-elementary: new recipe; python bindings for the elementary widget set (EFL) Jan 06 08:37:13 03Michael 'Mickey' Lauer  07org.openembedded.dev * r142f62a8d3 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@melo.nslu2-linux.org:openembedded into org.openembedded.dev Jan 06 08:37:16 03John Lee  07org.openembedded.dev * r6a221898b0 10openembedded.git/packages/netbase/netbase/ (om-gta01/interfaces om-gta02/interfaces): netbase: set the default nameserver to OpenDNS for om-gta0? Jan 06 08:37:19 03I-Fan, Chen  07org.openembedded.dev * r83dadf415e 10openembedded.git/packages/ (13 files in 5 dirs): Various recipes: add more detailed descriptions. Jan 06 08:37:24 03Erin Yueh  07org.openembedded.dev * ra8832b7664 10openembedded.git/packages/obexpush/ (files/opd_args obexpush_1.0.0.bb): Jan 06 08:37:29 obexpush_1.0.0.bb: add opd_args file to append arguments for starting Jan 06 08:37:31 opd Jan 06 08:37:33 03Holger Hans Peter Freyther  07org.openembedded.dev * rfa4f0bb849 10openembedded.git/packages/ipkg-utils/ (3 files in 2 dirs): Jan 06 08:37:36 ipkg-utils: Make it work on fedora 9. Jan 06 08:37:38 "control.tar.gz" got listed as "`control.tar.gz" The root cause is Jan 06 08:37:40 somewhere else in the parser. So this is a workaround that will help Jan 06 08:37:42 to find the right file. opkg on the device should use the proper ar Jan 06 08:37:44 executable and that should be able to extract the right files. Jan 06 08:37:46 03Graeme Gregory  07org.openembedded.dev * red3bee612c 10openembedded.git/packages/detect-stylus/ (detect-stylus_svn.bb files/install-detect-tsdevice.patch): detect-stylus_svn.bb: install and package detect-tsdevice Jan 06 08:37:52 03Koen Kooi  07org.openembedded.dev * r5f2c2447fe 10openembedded.git/ (conf/checksums.ini packages/xorg-lib/pixman_0.13.2.bb): pixman: add 0.13.2, disabled by default Jan 06 08:37:59 03Koen Kooi  07org.openembedded.dev * r0663295937 10openembedded.git/packages/angstrom/ (12 files in 2 dirs): Jan 06 08:38:02 e-wm-config-angstrom: e-wm profile for angstrom Jan 06 08:38:06 * http://www.flickr.com/photos/koenkooi/3169580059/ Jan 06 08:38:08 03Koen Kooi  07org.openembedded.dev * r9d2e4a220d 10openembedded.git/packages/tasks/task-beagleboard-demo.bb: task-beagleboard-demo: use angstrom e17 profile Jan 06 08:38:11 03Koen Kooi  07org.openembedded.dev * r56824421a2 10openembedded.git/ (conf/checksums.ini packages/gpe-conf/gpe-conf_0.2.8.bb): gpe-conf: add 0.2.8 Jan 06 08:38:14 03Koen Kooi  07org.openembedded.dev * r6613cff9f7 10openembedded.git/ (conf/checksums.ini packages/palantir/palantir_2.7.bb): palantier: add 2.7 Jan 06 08:38:17 03Koen Kooi  07org.openembedded.dev * rac5ca6c9bc 10openembedded.git/packages/dsplink/dmai_1.16.bb: dmai: fix description Jan 06 08:38:22 03Koen Kooi  07org.openembedded.dev * rb16036af1b 10openembedded.git/packages/dsplink/dmai_1.16.bb: Jan 06 08:38:27 dmai 1.16: add WIP recipe for CE DMAI driver Jan 06 08:38:29 * currently supports davinci, omap will be added later Jan 06 08:38:31 03Koen Kooi  07org.openembedded.dev * rdf824bb1c4 10openembedded.git/contrib/angstrom/build-feeds.sh: angstrom feed builder: add motion and palantir Jan 06 08:38:36 03Michael 'Mickey' Lauer  07org.openembedded.dev * r3044e9b7e1 10openembedded.git/ (2 files in 2 dirs): Jan 06 08:38:39 sane-srcrevs.inc: bump linux-openmoko-devel to a known good state Jan 06 08:38:41 linux-openmoko-devel: 2.6.28 works quite well now, will make this the new default soon Jan 06 08:38:43 03Koen Kooi  07org.openembedded.dev * rb090d9d86a 10openembedded.git/packages/linux/ (4 files in 3 dirs): linux-omap 2.6.28,git: fix overlay colourspace definition in DSS2 patches Jan 06 08:38:48 03Koen Kooi  07org.openembedded.dev * racb9402cea 10openembedded.git/packages/powervr-drivers/ (3 files): libgles-omap3: add 3...5 release, rework packaging a bit Jan 06 08:54:53 morning Jan 06 09:01:56 hi there, is there a way to specify a local source tree in SRC_URI within a recipe? Jan 06 09:02:29 I have a local overlay of a package, and I want it to use my own cloned git repository rather than the upstream one. Jan 06 09:04:46 atin_: use file:// Jan 06 09:05:05 can I specify the top of the source tree? Jan 06 09:05:35 garren|work: with the file:// Jan 06 09:05:46 not to sure but it might work Jan 06 09:06:15 garren|work: I actually tried that - it didn't. I finally had to fake my own "public" git repository. Jan 06 09:06:29 but I still had to commit changes - I just want it to use my working copy. Jan 06 09:08:01 ok... this is a little above my knowledge cant really help you. I havn't tried this before sorry Jan 06 09:08:44 garren|work: ok, thanks. it seems most of the people I know end up building stuff by hand instead of using OE unless just using the packages as they are already in OE. Jan 06 09:10:07 garren|work: so if you are actually creating patches to existing packages, you have to use something other than OE to do your work. I am really hoping there is some way to do this in OE. Jan 06 09:10:29 maybe I'll ask again when it is daytime in the US :) Jan 06 09:10:42 no create a local dircetory with your changes Jan 06 09:11:13 just finding you a link give me a sec :-) Jan 06 09:12:07 I think this is what you looking for if I understand you correctly http://elinux.org/BeagleBoardOpenEmbeddedDevelopment Jan 06 09:12:14 this is used for the beagleboard development Jan 06 09:12:23 garren|work: I wrote that ;) Jan 06 09:12:33 lol Jan 06 09:12:54 ok then I definitely cant help you, maybe koen can help... Jan 06 09:13:51 I want something more than what I put there - that shows how to "almost" get what I want :) Jan 06 09:14:14 ok, if you get the answer could you let me know Jan 06 09:14:30 one step is missing - the SRC_URI specification that would point at my local source tree rather than the official SRC_URI git repository. Jan 06 09:14:42 it might not be possible, Jan 06 09:17:33 I have started faking it by changing S= to my local tree, but hmm... just thought of something. I could modify S= and use bitbake -c do_compile Jan 06 09:19:52 nah. dang. oh well - goodnight :) Jan 06 10:55:25 hi. I have a recipe to submit, could someone review it before it goes to the bugzilla ? http://pastebin.com/m6d9a500 Jan 06 10:55:45 gm Jan 06 10:56:40 Longfield: there's no description field, and some other field(s) might be missing Jan 06 10:59:23 likewise: thx, I will fix this Jan 06 11:14:37 * broonie stabs bitbake Jan 06 11:22:42 Longfield: DEPENDS = not += Jan 06 11:24:28 thx hrw Jan 06 11:26:24 Longfield: your package does not generate lib*.so.* files? Jan 06 11:26:37 just lib*.so ones? Jan 06 11:31:11 yeah just lib*.so Jan 06 11:32:25 it's a small lib still in development here, so it's not really usefull to have "versions" of these libraries Jan 06 11:43:46 good morning Jan 06 11:46:41 hi all Jan 06 11:50:22 hi pb_ Jan 06 11:57:30 03Jan Lübbe  07org.openembedded.dev * r15fbbf158b 10openembedded.git/packages/tasks/task-fso-compliance.bb: task-fso-compliance: add more zoneinfo files for setting timezone via GSM Jan 06 12:04:31 03Jeremy Lainé  07org.openembedded.dev * re6bf92a777 10openembedded.git/packages/linux/ (4 files in 2 dirs): linux-2.6.27: update keyboard and LCD patches for boc01 Jan 06 12:16:23 whats the adress to the git repository? Jan 06 12:16:50 which one? Jan 06 12:16:54 git.openembedded.org? Jan 06 12:32:21 mickey|bs: s/.org/.net you mean? Jan 06 12:32:49 err, right Jan 06 12:35:09 ~curse our domain manager(s) Jan 06 12:35:10 May the fleas of a thousand camels infest your most sensitive regions, our domain manager(s) ! Jan 06 12:38:49 jeremy_laine: hey, long time no "see". So you got your own 8313 machine? :-) Jan 06 12:39:01 likewise: yup Jan 06 12:39:07 it's starting to shape up nicely Jan 06 12:39:13 mickey|bs: opinions on the logo? Jan 06 12:39:26 jeremy_laine: nice, may I ask (privately or publicly) what it does? Jan 06 12:39:58 likewise: it's a WiMAX residential gateway + mini service platform Jan 06 12:39:58 jeremy_laine: We're probably using the 15 as a cheap successor to the Intel XScale (includes GbE and PCIe) Jan 06 12:40:15 jeremy_laine: with capsense?! Jan 06 12:40:38 likewise: heh, it's not your vanilla gateway, admitedly Jan 06 12:40:48 jeremy_laine: sounds like a fun project then! Jan 06 12:41:10 I'm about to flash my u-boot from Linux (fingers crossed, no BDI nearby) Jan 06 12:41:40 likewise: there's a strong emphasis on web services, so a display + a couple of key were required Jan 06 12:42:49 mickey, the OOM problem was likely bad karma, due to OOM discussion in this channel ... Jan 06 12:44:54 hehe Jan 06 12:45:15 likewise: i'll comment soon Jan 06 12:48:19 mickey|bs: Thanks, it would be nice for the core team to act on it. Jan 06 12:48:54 servers probably should have OOM killer turned off Jan 06 12:53:57 good idea, how do i do that? Jan 06 12:57:19 XorA: and the hardware watchdog turned on! (Assuming virtual machine do support this) Jan 06 12:58:02 mickey|bs: no idea, I think it used to be in /sys somewhere Jan 06 12:58:05 mickey|bs: read about /proc/sys/vm/overcommit* options Jan 06 12:58:30 likewise: my VM has a watchdog, currently turned off because I like to see the serial console for errors Jan 06 13:00:13 bye Jan 06 13:14:16 hi mickey|bs Jan 06 13:14:24 heh Jan 06 13:19:05 * * OE Bug 4960 has been created by valentin.longchamp(AT)epfl.ch Jan 06 13:19:07 * * new package: dashel stream library Jan 06 13:19:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4960 Jan 06 13:22:06 * * OE Bug 4961 has been created by valentin.longchamp(AT)epfl.ch Jan 06 13:22:08 * * new package: aseba distributed robots control library Jan 06 13:22:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4961 Jan 06 13:22:42 if someone could have a look a the two abover bugs and if they are OK commit them, it would be very nice, thanks Jan 06 13:29:50 03Koen Kooi  07org.openembedded.dev * rbf6104e286 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Jan 06 13:30:01 03Jeremy Lainé  07org.openembedded.dev * rc228a100d3 10openembedded.git/packages/linux/ (4 files in 2 dirs): linux-2.6.27: update keyboard and LCD patches for boc01 Jan 06 13:34:13 mickey, I assume the old mtn keys don Jan 06 13:34:17 dont work anymore? Jan 06 13:35:41 mickey, did you see Harald's talk on gsm? Jan 06 13:47:07 * * OE Bug 4962 has been created by likewise(AT)gmx.net Jan 06 13:47:09 * * udhcpd deconfigures the network interface underneath a NFS rootfs Jan 06 13:47:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4962 Jan 06 13:51:26 03Leon Woestenberg  07org.openembedded.dev * r3ff023a25b 10openembedded.git/packages/busybox/busybox-1.11.3/udhcpscript.patch: busybox-1.11.3: udhcpd script: Do not deconfig network interface iff rootfs on nfs. Jan 06 13:51:29 03Leon Woestenberg  07org.openembedded.dev * rc03e405d58 10openembedded.git/packages/oprofile/ (5 files): oprofile: Include powerpc architecture files, due to powerpc/ppc namespace change. Jan 06 13:54:07 * * OE Bug 4962 has been RESOLVED (FIXED) by likewise(AT)gmx.net Jan 06 13:54:09 * * udhcpd deconfigures the network interface underneath a NFS rootfs Jan 06 13:54:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4962 Jan 06 14:02:24 Can someone take a look on the bitbake patch at http://pastebin.com/f71fef8b9 ? Jan 06 14:02:45 It solves a GIT fetcher issue we're having and I'd like to someone to push it Jan 06 14:05:42 otavio: if not ud.tag will match when tag = None? Jan 06 14:06:01 otavio: "if not ud.tag" is true when "tag = None"? Jan 06 14:06:42 likewise: yes Jan 06 14:07:04 otavio: I'm trying to read what it does. Can you explain what it solves? Jan 06 14:07:48 likewise: when you don't pass a tag, it uses master Jan 06 14:08:05 likewise: otherwise it fails since it find tag as "1" Jan 06 14:08:27 otherwise is when you do pass a tag? Jan 06 14:10:58 likewise: parse error; can you rephrase it? Jan 06 14:12:09 otavio: above, you say "when you don't pass a tag, it used master," "otherwise it fails". Otherwise is when you do pass a tag? Jan 06 14:12:45 likewise: it was not using master when I don't provide a tag; this patch fixes it Jan 06 14:13:53 otavio: ok Jan 06 14:15:27 otavio: I have seen that behaviour here as well, not sure if there is a specific reason for the current behaviour. I would post it on the OE and bitbake mailing list to raise attention Jan 06 14:15:28 likewise: :-) Jan 06 14:15:42 likewise: create Jan 06 14:15:45 likewise: I'll do it Jan 06 14:18:14 likewise: done; I mailed OE ml since I don't know the bitbake one. People can forward it there if required :-) Jan 06 14:45:03 Why again "u-boot-mkimage-openmoko-native" instead of just "u-boot-native" ? Jan 06 14:46:08 heh Jan 06 14:46:13 I wonder that also .... Jan 06 14:47:12 03Koen Kooi  07org.openembedded.dev * r62a428f026 10openembedded.git/packages/linux/ (3 files in 3 dirs): linux-omap 2.6.27: sync powervr code with latest code drop from TI Jan 06 14:49:17 03Koen Kooi  07org.openembedded.dev * re625b67830 10openembedded.git/packages/linux/linux-omap_2.6.27.bb: linux-omap 2.6.27: bump PR Jan 06 15:09:20 anyone - is there *anyway* in a bitbake recipe (local overlay recipe actually) to make SRC_URI point to the top of a local source tree as opposed to a git://, cvs://, svn:// or tar files etc? Jan 06 15:10:27 file:// Jan 06 15:10:41 that doesn't take the top of a source tree. it takes lists of sources Jan 06 15:10:45 files. Jan 06 15:10:59 oh, you can just point ${S} to where you want I think Jan 06 15:11:12 I want to point it at /home/atin/oe/local/sources/u-boot-omap3-dev Jan 06 15:11:55 XorA - yes, I discovered that, but it requires a clean everytime otherwise bitbake thinks the source it downloaded from SRC_URI was already compiled and it doesn't need to do anything. Jan 06 15:12:43 atin_: bitbake -c compile -f Jan 06 15:12:52 -f forces it to re-run compile Jan 06 15:13:37 XorA: ah! I was looking for a -f last night but maybe it was 4am and my brain just skipped it! Jan 06 15:52:24 XorA: cool! that did it. Jan 06 15:58:44 re Jan 06 16:10:17 03Leon Woestenberg  07org.openembedded.dev * r7a9edc3215 10openembedded.git/ (11 files in 2 dirs): dtc: Add 1.2.0 and use -native.inc and .inc files. Jan 06 16:15:35 03Koen Kooi  07org.openembedded.dev * r993b786f90 10openembedded.git/packages/angstrom/ (3 files in 2 dirs): e-wm-config-angstrom: add some tweaks Jan 06 16:59:35 hello,for the libsdl-mixer should we force the inclusion of vorbis(with deps or --flags)? because if you build it like this without building vorbis before it fails to include the vorbis support Jan 06 17:34:05 * * OE Bug 4963 has been created by chrisfriedt(AT)gmail.com Jan 06 17:34:07 * * package.bbclass - uncaught exception from os.stat() on symbolic link Jan 06 17:34:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4963 Jan 06 18:20:06 * * OE Bug 4964 has been created by treitmayr(AT)devbase.at Jan 06 18:20:08 * * supertux: Ported fixed-point patch from 0.1.2 to 0.1.3 Jan 06 18:20:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4964 Jan 06 18:40:01 strange... I find that bitbake 1.8.10 expands shell vars in included files (ie build/local.conf) but 1.8.11 does not - how come? Jan 06 18:45:15 03Koen Kooi  07org.openembedded.dev * r785b8609e6 10openembedded.git/packages/powervr-drivers/omap3-sgx-modules_1.3.13.1397.bb: omap3-sgx-modules: recipe to build sgx modules using the kbuild system from the GPL'ed imgtec sources Jan 06 18:45:17 without shell vars expanding your conf files are a mess - how are those of you using bitbake 1.8.11 handling something like 'BBFILES="${OP_TOP}/packages/*/*.bb" ? Jan 06 18:47:08 03Koen Kooi  07org.openembedded.dev * r665c67013b 10openembedded.git/packages/linux/ (4 files in 2 dirs): linux-omap 2.6.27: remove outdated PVR patch since we have external modules now Jan 06 18:47:38 03Koen Kooi  07org.openembedded.dev * re7dd055610 10openembedded.git/classes/module.bbclass: module.bbclass: pass KERNELDIR as well Jan 06 18:52:15 tharvey: BB_ENV_EXTRAWHITE Jan 06 18:53:44 hello, i've done a RDEPEND for perl packages in the frozen-bubble_2.2.0 recipe: http://pastebin.com/m4ab26f6b but when I install the ipk...the perl deps aren't installed!!! what should i change to the recipe? Jan 06 18:57:18 I'm getting an error during the staging of libusb1: /bin/sh: /home/srussell/embedded/oe/build/tmp/staging/x86_64-linux/usr/bin/mkdir: No such file or directory. when is the native mkdir supposed to be created? I suppose I could just manually symlink it to my host version, but I'd rather fix the problem. thanks! Jan 06 18:57:20 jo Jan 06 18:57:26 woglinde, hello Jan 06 18:57:30 hi gnutoo Jan 06 18:57:47 woglinde, how can i make a package depend on perl modules? Jan 06 18:58:01 woglinde, didn't have time to send mail sorry Jan 06 18:58:16 gnutoo DEPENDS = perl Jan 06 18:58:26 RDEPENDS = perl-moofoo Jan 06 18:58:38 but this is discribed in the manual Jan 06 18:58:46 yes but it didn't work for me Jan 06 18:59:20 hm why? Jan 06 18:59:21 http://pastebin.com/m4ab26f6b Jan 06 18:59:25 whats going wrong? Jan 06 18:59:36 when i do opkg insatll package it only pull the following deps: Jan 06 18:59:41 tharvey: read the release announcement? Jan 06 18:59:58 libsdl-mixer-1.2-0 libsdl-pango1 Jan 06 19:00:18 gnutoo whats the error? Jan 06 19:00:24 woglinde, no error Jan 06 19:00:32 woglinde, just that it doesn't pull the deps Jan 06 19:00:37 tharvey: env. variables like SILENT, DEBUG, MACHINE, V, Q... break builds of various buildsystems... now we have sanitized the environment Jan 06 19:01:02 mabe i'll analyze the ipk Jan 06 19:05:05 the ipk: http://rafb.net/p/VZbrlH90.html Jan 06 19:09:21 woglinde, so what should i do now? i bet the deps name are correct otherwise i would have an error Jan 06 19:10:44 i'll grep for RDEPEND on the recipies Jan 06 19:11:03 hmm, it seems like somewhere along the line an extra "/tmp/staging/x86_64-linux" got created in "BUILDDIR/tmp/staging/x86_64-linux" and that's where these binaries got installed. So their path is "BUILDDIR/tmp/staging/x86_64-linux/tmp/staging/x86_64-linux/usr/bin/mkdir". Typo in the build script somewhere or is this where they're supposed to be? Jan 06 19:13:10 zecke, release notes for what? bitbake 1.8.11? I'll look there for more info Jan 06 19:14:14 I can definately understand sanitizing the env... I'm just trying to understand the right mechanism to do what I used to do with bitbake 1.8.10 which was to export some shell vars and allow the various bitbake conf files to inherit those vals Jan 06 19:14:22 have to go afk for a bit Jan 06 19:15:08 tharvey: see the oe-devel ml archive from the last days... whitelist the vars you want to inject Jan 06 19:16:45 grep did return some result but nothing interresting...there are packages like mine using perl-module-something... Jan 06 19:16:57 like ddclient for instance Jan 06 19:18:18 i'll bitbake it in order to see if it works Jan 06 19:24:18 it works Jan 06 19:24:35 i'll rebuild the frozen-bubble package with a higher rev Jan 06 19:27:08 doesn't change anything Jan 06 19:28:39 ah...mabe i get it Jan 06 19:28:49 i must do RDEPENDS_PN mabe Jan 06 19:29:11 i'm having trouble with sysvinit, i always get a kernel panic not syncing attempted to kill init with no other information. Is there something that I can do to get a real debug output? Jan 06 19:31:00 jo kergoth Jan 06 19:31:13 Gnutoo sorry without the whole recipe I cant tell you Jan 06 19:31:30 hey Jan 06 19:31:32 ok i'll try that and then i'll post the recipe if it doesn't work Jan 06 19:31:36 thanks a lot Jan 06 19:31:51 leonardo_: try to boot into /bin/sh Jan 06 19:31:58 Gnutoo ah yes Jan 06 19:32:06 data is non standard-package Jan 06 19:32:34 and for sure clashes with frozenbubble_version.bb Jan 06 19:32:48 Gnutoo but why the heck the -data package needs the depends? Jan 06 19:32:57 ah sorry Jan 06 19:32:59 okay Jan 06 19:33:07 its frozen-bubbles self Jan 06 19:33:08 woglinde, lol Jan 06 19:33:12 yes Jan 06 19:33:30 coulb please paste the whole recipe? Jan 06 19:33:44 i try it and then i paste it Jan 06 19:34:41 khem: I get the exact same thing using init=/bin/sh in bootargs. VFS: Mounted root (nfs filesystem) \ Freeing init memory: 124K \ Kernel panic - not syncing: Attempted to kill init! Jan 06 19:35:30 leonardo_: hmm are you booting over nfs ? Jan 06 19:35:40 yes Jan 06 19:35:46 ah Jan 06 19:35:55 leonardo_: then do the following Jan 06 19:36:11 Gnutoo and the name Jan 06 19:36:15 of the recipe Jan 06 19:36:16 the modules are now included Jan 06 19:36:27 ok ok i'll past the recipe before testing it Jan 06 19:36:29 http://rafb.net/p/Du7Olv94.html Jan 06 19:36:29 gnutoo hehe Jan 06 19:36:31 open etc/network/interfaces in your rfs Jan 06 19:36:58 khem: ok Jan 06 19:37:08 search for eth0 Jan 06 19:37:30 comment out the first occurance Jan 06 19:37:34 Gnutoo hm put the DESCRIPTIOn line as the first line Jan 06 19:37:38 the auto eth0? Jan 06 19:37:43 leonardo_: yes Jan 06 19:37:51 and the reboot Jan 06 19:37:56 ok Jan 06 19:37:57 thanks Jan 06 19:38:36 done Jan 06 19:39:28 khem: i get the same thing Jan 06 19:39:35 hmm...could the problem be that coreutils-native wasn't compiled before libusb1 tried to use it? Should coreutils-native be a dependency? Jan 06 19:40:38 leonardo_: then are you sure that your nfs server is setup correctly Jan 06 19:40:51 leonardo_: and kernel supports NFS Jan 06 19:41:24 Gnutoo hm does frozen bubble use gettext? Jan 06 19:41:35 yes it does...with perl Jan 06 19:41:43 khem: i believe so, wouldn't it have a problem and not say VFS: Mounted root (nfs filesystem) if it couldn't get to it? And wouldn't it say can't find init? Jan 06 19:41:45 directly or via perl? Jan 06 19:41:51 via liblocale-gettext-perl Jan 06 19:41:56 args Jan 06 19:42:00 what the hell is this Jan 06 19:42:11 hm okay Jan 06 19:42:16 leonardo_: and a dhcp server Jan 06 19:42:36 there are *.mo files for frozen-bubble so that are gettext files Jan 06 19:42:37 Gnutoo normaly a locales packages for each lang is generated Jan 06 19:42:45 it is Jan 06 19:42:52 FILES_${PN}-locale = " \ Jan 06 19:43:03 khem: as part of bootup it runs dhcp and I see that it gets an IP Jan 06 19:43:05 is then not right in this case Jan 06 19:43:46 leonardo_: if you have this working then it should atleast boot into /bin/sh IMO Jan 06 19:44:04 if not then your rfs is seriously hosed up Jan 06 19:44:17 Gnutoo can you build frozen bubble without locales? Jan 06 19:44:22 http://rafb.net/p/sVHd9C47.html Jan 06 19:44:27 khem: it's the rootfs generated by OE build Jan 06 19:44:30 don't know Jan 06 19:44:34 i'll look Jan 06 19:44:41 khem: gpe-image Jan 06 19:44:47 leonardo_: which distro ? Jan 06 19:44:52 khem: angstrom Jan 06 19:44:56 and is it from .dev ? Jan 06 19:45:05 for which arch ? Jan 06 19:45:07 khem: yes, did a git pull this morning Jan 06 19:45:15 khem: at91sam9261ek Jan 06 19:45:27 Gnutoo otherwise you have to honor USE_NLS Jan 06 19:45:28 woglinde, why do you want to build it without the locales? Jan 06 19:45:38 because of diskspace Jan 06 19:45:40 leonardo_: I usually boot over nfs on another arm platform have had no problems Jan 06 19:45:51 other than the one I mentioned to you Jan 06 19:45:52 woglinde, evry locale is in a separate package... Jan 06 19:45:54 and compile speed Jan 06 19:45:58 ah i understand Jan 06 19:46:15 because people disable locales at building Jan 06 19:46:28 khem: i haven't had any problems when I've created my own rfs, that's why I don't know what's going on here, also i'm using the 2.6.28 kernel from kernel.org Jan 06 19:46:30 so it would break Jan 06 19:46:34 gnutoo right Jan 06 19:46:38 hm it wouldnt break Jan 06 19:46:51 but would look a bit odd Jan 06 19:47:03 ok lol...that would need a bit of hacking... Jan 06 19:47:16 hm some near time in the future I will ask koen why glib needs locales all the time Jan 06 19:47:44 Gnutoo yes and because gettext isnt used directly you cant do inherit gettext Jan 06 19:48:42 khem: i'm using bootargs = console=ttyS0,115200n8 noinitrd rw ip=dhcp root=/dev/nfs nfsroot=XXX.XXX.XXX.XXX:/path/to/fs,no_lock init=/bin/sh Jan 06 19:52:09 jo raster Jan 06 19:52:12 and likewise Jan 06 19:52:36 leonardo_: what does exportfs say on your host ? Jan 06 19:53:38 khem: the top level directory /nfs_export where i put rootfs Jan 06 19:54:34 khem: /etc/exports says /nfs_export *(rw,sync,no_subtree_check) Jan 06 19:57:23 leonardo_: may be you could add no_root_squash Jan 06 19:57:29 so the options Jan 06 19:57:34 woglinde, so what should I do exactly? hack the Makefile to build only some locales? Jan 06 19:57:42 *(rw,sync,no_subtree_check,no_root_squash) Jan 06 19:57:48 s/to build only/to be able to build only/ Jan 06 19:58:16 gnutoo if the buildsystem dont allow it Jan 06 19:58:21 let it as it is Jan 06 19:58:30 this would be to much pain Jan 06 19:58:40 lol...here's the Makefile: http://rafb.net/p/cp9mz343.html Jan 06 19:58:57 dont they use autotools? Jan 06 19:59:04 no they use perl Jan 06 19:59:19 perl Malkefile.pl Jan 06 19:59:28 * khem thinks a choice between bad and worse :) Jan 06 19:59:54 Gnutoo okay hm let it as it is Jan 06 20:00:03 to much pain to gain this little effort Jan 06 20:00:05 basically there is a topdir makefile that calls perl Makefile.PL that creates the c makefile etc... Jan 06 20:00:06 ok Jan 06 20:00:19 so i will submit it in the bugreporting system...lol Jan 06 20:00:53 khem: i get the same kernel panic Jan 06 20:01:06 khem: i made sure to restart nfs-kernel-server Jan 06 20:02:16 leonardo_: and also nfs-common ( if you are on debian-like system) Jan 06 20:04:16 khem: ok, did nfs-common also and same result Jan 06 20:05:33 re Jan 06 20:05:43 jo florian Jan 06 20:05:53 khem: should exporting /nfs_export be enough or I have to export the subdirs too? Jan 06 20:06:05 leonardo_: this should be enough Jan 06 20:06:22 someone from Turkey is trying to hack my system Jan 06 20:07:45 khem: did everything look good in my bootargs? Jan 06 20:10:06 * * OE Bug 4965 has been created by GNUtoo(AT)no-log.org Jan 06 20:10:08 * * frozen-bubble update to version 2.2.0 and better package Jan 06 20:10:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4965 Jan 06 20:13:48 Elena are you awake? Jan 06 20:21:11 leonardo_: Enable user debugging in kernel may be there is some bug somewhere I dont know Jan 06 20:21:18 mickey, this weekends outage reminded of this : http://thedailywtf.com/Articles/ITAPPMONROBOT.aspx Jan 06 20:21:25 it works ok for me on my little osk Jan 06 20:22:11 khem: which option is that? Jan 06 20:23:04 add user_debug=-1 debug Jan 06 20:23:08 to bootargs Jan 06 20:23:21 but you need to enable it in kernel too Jan 06 20:23:37 has anyone used the jack recipe before? Jan 06 20:24:06 * * OE Bug 4966 has been created by GNUtoo(AT)no-log.org Jan 06 20:24:07 * * new bb recipe : sdlpango, needed for frozen-bubble_2.2.0 Jan 06 20:24:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4966 Jan 06 20:25:43 khem: i just booted to an old rootfs that I had done before using OE and everything worked fine. I put it in the /nfs_export directory so I think the host is setup correctly and the bootargs Jan 06 20:28:42 I've used the jack recipe but I made my own Jan 06 20:28:51 uh, actually I'm not sure Jan 06 20:29:05 * * OE Bug 4967 has been created by GNUtoo(AT)no-log.org Jan 06 20:29:07 * * libsdl-mixer lack libogg libvorbis as DEPENDS Jan 06 20:29:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4967 Jan 06 20:29:10 I may have just copied it from dev because it wasn't in stable(?) Jan 06 20:29:40 khem: I have something now, with that debug option init(1): undefined instruction: pc=00009e54 Jan 06 20:30:56 re Jan 06 20:31:04 hi hrw Jan 06 20:32:23 garren: I made a recipe for jack 0.109.2 and then I ended up not using jack because it didn't seem well behaved IIRC. I was trying to use it with portaudio. Jan 06 20:32:34 woglinde, do you have some time for hacking on boost_1.36? Jan 06 20:33:19 or should we do that another day? Jan 06 20:33:46 gnutto hm anotherday Jan 06 20:33:55 ok thanks Jan 06 20:34:07 and thanks for the advises on frozen-bubble Jan 06 20:34:42 Gnutoo yes no porb Jan 06 20:34:44 aeh prob Jan 06 20:34:45 archae0pteryx: ok because I need jack for another recipe I've written/testing and I cant find jackd in my rootfs Jan 06 20:35:01 Gnutoo you could try on boost alone Jan 06 20:35:17 gnutoo try the packagname variant as crofton sugessted Jan 06 20:35:19 woglinde, i've already tried and i'm blocked on the packaging Jan 06 20:35:51 woglinde, the problem is that the python code in the recipe seems not to be executed Jan 06 20:35:59 and it's python __anonymous() { Jan 06 20:36:08 hm Jan 06 20:36:13 try to search in other recipes Jan 06 20:36:17 for python code Jan 06 20:36:20 i've done that Jan 06 20:36:28 hm Jan 06 20:36:29 okay Jan 06 20:36:48 i even found better code in abiword-plugins-recipe but it didn't work either Jan 06 20:37:09 hm strange Jan 06 20:38:12 it puts evrything in -dev Jan 06 20:38:40 garren: you need to install jack-server and libjack Jan 06 20:38:49 even if there is no FILES_${PN}-dev Jan 06 20:38:53 inside the recipe Jan 06 20:39:24 garren: see PACKAGES in the recipe Jan 06 20:39:26 Gnutoo this handled in in other class Jan 06 20:39:28 archae0pteryx: are those not installed by the jack recipe? Jan 06 20:39:51 woglinde, ah ok mabe i should inherit something... Jan 06 20:40:28 garren: they are all built by the recipe but apparently they are split into separate packages in the FILES_* statements Jan 06 20:40:36 (I am inferring here) Jan 06 20:40:38 arch Jan 06 20:40:47 archae0pteryx: ok I'll try that thanks Jan 06 20:41:05 abiword has inherit autotools.... Jan 06 20:41:16 but boost uses boost's jam Jan 06 20:41:18 (bjam) Jan 06 20:41:38 03Florian Boor  07org.openembedded.dev * r85ecd0988b 10openembedded.git/packages/unicap/unicap_0.9.3.bb: unicap: fix packages Jan 06 20:41:39 03Florian Boor  07org.openembedded.dev * r77620ff555 10openembedded.git/conf/machine/htcuniversal.conf: htcuniversal.conf: Remove duplicated EXTRA_ARCHS entries. Jan 06 20:41:39 03Florian Boor  07org.openembedded.dev * r0e2f59b4fc 10openembedded.git/packages/unicap/unicap_0.9.4.bb: unnicap: Add 0.9.4 release Jan 06 20:41:40 03Florian Boor  07org.openembedded.dev * red1e1b42c0 10openembedded.git/packages/unicap/ (3 files in 3 dirs): ucview: Fix compiling and move patches around Jan 06 20:41:44 03Florian Boor  07org.openembedded.dev * rfbbadf9f2a 10openembedded.git/packages/gpe-conf/gpe-conf_svn.bb: gpe-conf: Update svn recipe to new version and use of SRCREV. Jan 06 20:41:46 03Florian Boor  07org.openembedded.dev * rdac0500fea 10openembedded.git/conf/distro/include/sane-srcrevs.inc: saane-srcrevs.inc: Add entry for gpe-conf Jan 06 20:41:51 03Florian Boor  07org.openembedded.dev * r642a84be4a 10openembedded.git/packages/meta/meta-toolchain-gpe.bb: meta-toolchain-gpe: Remove GPE Phone Edition task. Jan 06 20:41:54 03Florian Boor  07org.openembedded.dev * r9951a8394f 10openembedded.git/packages/tasks/task-sdk-base.bb: task-sdk-base: Fix dependencies Jan 06 20:41:57 03Florian Boor  07org.openembedded.dev * rb107519967 10openembedded.git/packages/tasks/task-gpe-connectivity.bb: task-gpe-connectivity: Use gpe-mini-browser2 instead of old one. Jan 06 20:42:31 Gnutoo packages-bla_files default is handled at bitbake.conf Jan 06 20:42:40 in the conf-dir Jan 06 20:42:48 woglinde, ah ok thanks a lot!!! Jan 06 20:44:37 archae0pteryx: it tells me that ubuildable dependancy chain for libjack and jack-server Jan 06 20:48:58 garren: I'm using the last stable, not sure if that helps Jan 06 20:49:47 garren: looks like I also have alsa-plugins 1.0.15 in my local recipes Jan 06 20:50:01 I think I solved it I put jack-server in my RDEPENDS of my recipe Jan 06 20:50:06 and that should be the end of the dependency chain Jan 06 20:50:11 ok Jan 06 20:50:33 how did you do it? also only RDEPENDS? Jan 06 20:51:02 garren: I don't think I built jack-server directly, just jack Jan 06 20:51:51 I'm trying to use OE using the angstrom distro. When i boot to the rootfs I get "init (1): undefined instruction: pc=00009e54" after the mounting of the rootfs when it tries to run init. Does anyone know why I would get that? Jan 06 20:51:52 it's been a couple of months since I've touched it Jan 06 20:53:00 archae0pteryx: cool thanks for the help... I'll keep going and see what happens Jan 06 20:53:20 leonardo: I am ignorant but I would first check the architecture being used to build against my CPU Jan 06 20:53:41 the options to gcc Jan 06 20:54:59 can also use the command 'file' to check /sbin/init Jan 06 20:56:19 archae0pteryx: i've used file and gotten ELF 32-bit executable, ARM, version 1 (SYSV) ......dynamically linked Jan 06 20:56:26 i'm using an ARM9 processor Jan 06 20:56:29 at91sam9261 Jan 06 20:57:25 is it ok for the init to be dynamically linked? Jan 06 20:57:33 is there a way to tell what libraries it needs? Jan 06 20:57:44 my init is dynamically linked too Jan 06 20:57:47 ok Jan 06 20:58:16 i put at91sam9261ek in the local.conf for OE Jan 06 20:58:27 so i think it should compile to the right machine Jan 06 20:58:31 Dunno about those Jan 06 20:59:17 My init says ELF 32-bit LSB executable, ARM, version 1 (SYSV), for GNU/Linux 2.6.14, dynamically linked (uses shared libs), stripped Jan 06 20:59:51 exactly the same Jan 06 21:00:00 Yours doesn't have LSB Jan 06 21:00:04 Does it? Jan 06 21:00:09 leonardo_: ldd binary displays libraries Jan 06 21:00:20 sorry it does Jan 06 21:00:41 using two computers so i'm not able to copy/paste Jan 06 21:01:05 actually, my /sbin/init is a link to /sbin/init.sysvinit Jan 06 21:01:11 so is mine Jan 06 21:01:19 but init.sysvinit is the same as what you pasted Jan 06 21:01:22 it was just an idea not sure if thats it Jan 06 21:01:42 what processor are you using? Jan 06 21:02:02 pxa270 Jan 06 21:03:43 i guess there is a bug in creating the cross compiler for my arch? Jan 06 21:04:34 That's not the first conclusion I would jump to :) are there prebuilt binaries to try somewhere for that machine Jan 06 21:04:41 the cross compiler is in OE/tmp/cross/armv5te/bin Jan 06 21:05:14 archae0pteryx: i have a very simple rootfs that i made with ELDK cross compiler that works correctly, so i have binaries in that Jan 06 21:06:10 I use angstrom and my cc isn't in the same place Jan 06 21:06:26 i was trying to use angstrom also Jan 06 21:06:30 hmm Jan 06 21:06:38 where is your's? Jan 06 21:07:17 tmp/cross/arm-angstrom-linux-gnueabi Jan 06 21:07:58 hmm, i don't have that directory, just armv5te Jan 06 21:08:17 using angstrom-2007.1 Jan 06 21:08:21 (I am) Jan 06 21:08:43 can you set up eldk to use the OE compiler Jan 06 21:08:53 well, i have tmp/cross/armv5te/arm-angstrom-linux-gnueabi Jan 06 21:09:14 I'm using stable, maybe that's changed in more recent versions? Jan 06 21:09:17 using angstrom-2008.1 Jan 06 21:09:21 probably Jan 06 21:09:53 well if you could eldk to work with the OE compiler and the same gcc options that would eliminate one variable Jan 06 21:10:00 at least you have something that works :) Jan 06 21:10:19 yeah, just don't have all the packages that angstrom would give me Jan 06 21:11:25 sure but if you are just trying to isolate the problem Jan 06 21:19:16 mabe i have forgetten to add the PACKAGES_DYNAMIC variable...will retry with it Jan 06 21:57:39 woglinde, i must go...bye...by the way could someone commit for me frozen-bubble and xboard? Jan 06 21:58:57 hm okay will do it Jan 06 21:59:08 thanks a lot Jan 06 22:00:41 bye Jan 06 22:43:16 gm raster Jan 06 22:44:35 woglinde: mooornin Jan 06 22:50:02 03woglinde  07org.openembedded.dev * r2ff9ef21e5 10openembedded.git/packages/frozen-bubble/ (Makefile.PL.patch Makefile.patch Makefile_top.patch): frozen-bubble: put the patches into their own dir Jan 06 22:55:35 why theheck the system wants gcc-native Jan 06 23:02:15 has there been a recent change that no entry in conf/checksums.ini is an ERROR now vs a NOTE? Jan 06 23:02:23 yes Jan 06 23:02:32 its must now Jan 06 23:04:28 woglinde: add it to ASSUME_PROVIDED Jan 06 23:05:04 khem was only with glibc Jan 06 23:05:12 woglinde: are you using qemu ? Jan 06 23:05:37 IIRC qemu needs it Jan 06 23:05:40 hm Jan 06 23:05:42 right Jan 06 23:06:00 qemu can not be compiled with gcc-4.x yet Jan 06 23:06:05 khem I know this Jan 06 23:06:10 ah... ok - so what should I do for source packages from a local overlay? can I somehow add to the list of files checked for checksums? (this is a bitbake change/feature right?) Jan 06 23:06:28 conf/checksum.ini Jan 06 23:10:02 ah ok... all conf/checksum.ini's in BBPATH... Jan 06 23:49:06 03Koen Kooi  07org.openembedded.dev * r4f6da86637 10openembedded.git/conf/machine/ (beagleboard.conf omap3-pandora.conf omap3evm.conf): omap3 machines: add omap3-sgx-modules to RECOMMENDS Jan 06 23:49:12 03Koen Kooi  07org.openembedded.dev * r9b1666b54a 10openembedded.git/packages/tasks/task-base.bb: task-base: bump PR Jan 06 23:54:03 * eekee wonders whether to try building a uclibc-based system for his spitz Jan 06 23:59:18 what's IMAGE_FSTYPES? What are "the required image file system types"? Jan 07 00:01:43 eekee: Depends on what you want to use for your device. For the Spitz jffs2 might be what you want. Jan 07 00:01:50 good night Jan 07 00:01:56 um, night Jan 07 00:03:09 eekee: how are you going to get your RFS onto your device Jan 07 00:04:08 archae0pteryx: I'm planning on installing to the internal drive. Unpacking tar files on the device would be fine Jan 07 00:05:02 I guess with the jffs2 option I get a filesystem I can put on a flash card? I thought jffs2 wasn't for flash cards Jan 07 00:05:18 is it mtd or sd/cf Jan 07 00:05:32 former you want jffs2 to burn from boot loader Jan 07 00:06:30 ahh, it's an internal cf drive, but I dont' know where the kernel is Jan 07 00:06:55 I think I can supply a kernel file gzipped & the updater will flash that Jan 07 00:08:39 AFAIK kernel doesn't have to do with the option you described Jan 07 00:08:44 ok good Jan 07 00:19:58 do I need a cross-compiler? Or will bitbake take care of that? Jan 07 00:20:28 oh... Jan 07 00:20:32 archae0pteryx: what's an RFS? Jan 07 00:20:40 root file system Jan 07 00:21:02 ah cheers, yeah I can instlal that with a tar file. Jan 07 00:21:06 *install Jan 07 00:21:23 IIRC the cross compiler builds when you do anything Jan 07 00:21:28 coolio Jan 07 00:28:19 mjello Jan 07 00:31:51 * eekee pokes Zta's word Jan 07 00:32:40 Be my guest. That word needs some poking. Jan 07 00:32:51 yeah hehe Jan 07 00:34:21 I'm creating my-own-image.bb and want to install some packages on the image. Should I use IMAGE_INSTALL or PACKAGE_INSTALL ? What's the difference? Jan 07 00:41:06 zta: i use image_install, dunno about package_install Jan 07 00:41:42 archae0pteryx, yeah... http://bec-systems.com/oe/html/commonuse_new_image.html Jan 07 00:42:25 never seen that Jan 07 00:42:29 image_install works at least Jan 07 00:42:33 unless they've changed it Jan 07 00:42:35 i'm on stable Jan 07 00:44:37 I'm using image_install too. It works. Jan 07 00:44:39 Just wondering. Jan 07 00:47:00 If you find out tell me ;-) Jan 07 02:13:41 weird.. invoking the "populate_staging" works, while "stage" fails.. Jan 07 02:14:04 bitbake -b -c stage -f -D **** ENDING LOGGING AT Wed Jan 07 02:59:57 2009