**** BEGIN LOGGING AT Tue May 03 02:59:58 2011 May 03 08:04:00 do we by now have a global way to scmdata=keep or should i cook up something? May 03 08:04:51 blindvt`: default is off add it to SRC_URI where needed (done for efl) May 03 08:05:23 yea, i know but i find it more convenient to be able to control that in local.conf once and for all May 03 08:06:10 I don't see such use-case May 03 08:06:37 if some recipe really needs scmdata then it's worth adding to recipe and fixing it for everybody May 03 08:06:46 JaMa|Wrk: i only need to keep the data if i have to fix stuff (so i can conveniently make patches and/or apply it directly. Normal op is not to keep it, of course May 03 08:07:21 why don't you cook patches in checkout in downloads dir? May 03 08:08:23 hm.. probably because i'm used to just step into the work dir where i can directly build and runtime-test it, i think May 03 08:13:26 hmz.. fetch2 still is not as generalized as it should be, e.g. with scmdata handling and more general all entities used for tar. I had a patch somewhere to that effect May 03 10:37:35 hi all how can i compile kernel headers using oe? May 03 10:38:03 strange question May 03 10:38:23 why would you need to do that exactly May 03 10:38:28 I don't understand.... May 03 10:38:47 you can bitbake kenrel headers, that would stage tehm May 03 10:39:00 then with that you can build the libc May 03 10:39:25 but you don't compile headers, at least not without .c May 03 10:39:42 GNUtoo, for some drivers May 03 10:39:54 ah ok May 03 10:40:04 look at the modules recipes then May 03 10:40:52 for instance compat-wireless May 03 10:41:21 KLIB_BUILD=${STAGING_KERNEL_DIR} May 03 10:41:24 that for instance May 03 10:41:29 in compat-wireless.inc May 03 10:41:40 and: May 03 10:41:50 inherit module May 03 10:42:25 no need of the headers, use the full kernel source May 03 10:46:51 ok May 03 10:48:42 hi guys May 03 10:49:26 Can I define pkg_postinst_* scripts in my custom image.bb file? May 03 10:49:46 for example pkg_postinst_qt4-embedded ? May 03 10:56:31 i have compiled the kernel 2.6.37 for beagle board it is working fine but it shows backtrace msg like http://pastebin.com/3QzfAhzp May 03 10:56:39 compiled kernel with oe May 03 11:07:45 aarti, it's not oe but the kernel itself May 03 11:09:37 GNUtoo, ya i knw this is cmg frm kernel only May 03 11:10:12 you could just ignore the warning May 03 11:10:15 or try to fix it May 03 11:10:25 or bugreport it to the kernel that oe is using May 03 11:10:46 GNUtoo, ok thnx May 03 11:13:55 if you choose to try to deal with it look at the recipe of your kenrel May 03 11:14:02 there should be an SRC_URI inside May 03 11:14:55 ya m kernel is comming frm git May 03 11:15:13 git.kernel.org May 03 11:19:53 ok May 03 11:19:58 no additional patches? May 03 11:20:08 what is your kernel recipe? May 03 12:47:18 Is there someone who knows how to install only qt4-embedded libraries without installing qt docs, examples, tools? Do I need to modify the receipe or there are some other methods? May 03 12:47:58 mrAlmond: split packages are produced May 03 12:48:29 I presume you are just installing the qt4-embedded package right now? that's the meta package that pulls in everything, you probably don't want that May 03 12:48:36 yes May 03 12:48:53 the rootfs grows a lot May 03 12:49:05 the qtopia doc directory is more or less 250 MB May 03 12:49:11 you can have a look at the packages-split subdir of the qt4-embedded workdir to see the packages produced May 03 12:49:30 great! May 03 12:50:31 Am I supposed to find this packages-split directory in meta-openembedded/meta-oe/recipes-qt/qt4/qt4-embedded ? May 03 12:50:55 because I can't find it May 03 12:51:04 mrAlmond: build/tmp/work//qt4-embedded May 03 12:51:08 or something along those lines May 03 12:51:13 ookk May 03 12:51:17 tnx!! May 03 12:52:47 So in my custom image file I don't have to specify IMAGE_INSTALL += "qt4-embedded" but a list of these split packages? May 03 12:53:05 yes, whatever packages you want/need May 03 12:53:17 ok thank you! May 03 13:27:06 Hi all, little problem with tzcode : May 03 13:27:08 NOTE: Running task 25 of 73 (ID: 8, /home/guilig/dev/OE/openembedded/recipes/tzcode/tzcode-native_2011g.bb, do_fetch) May 03 13:27:08 NOTE: package tzcode-native-2011g-r4.0: task do_fetch: Started May 03 13:27:08 ERROR: tzcode-native-2011g: ftp://elsie.nci.nih.gov/pub/tzdata2011g.tar.gz cannot check archive integrity May 03 13:27:08 NOTE: package tzcode-native-2011g-r4.0: task base_do_fetch: Failed May 03 13:27:08 ERROR: Function 'base_do_fetch' failed May 03 13:27:19 any suggestions ? May 03 13:29:07 what is the policy about packages in oe-core meta-demoapps? May 03 13:29:34 it has fuse but it is outdated and I wish to update it. Currently I did it into meta-oe but seems more logical to avoid duplicating meta-data May 03 13:39:59 bluelightning : I've done a list of all the packages_split I need but when I start buildind the image after I while I get errors like this : "opkg_install_cmd: Cannot install package qt4-embedded-common" May 03 13:40:23 are the packages_split name something I must insert in IMAGE_INSTALL ? May 03 13:41:29 mrAlmond: it should be OK to put the names under packages_split into IMAGE_INSTALL May 03 13:45:24 because it seems that packages names under packages_split are not recognized when passed in IMAGE_INSTALL May 03 13:45:35 "opkg_install_cmd: Cannot install package qt4-embedded-common" May 03 13:46:35 mrAlmond: it should work. For example, I have this (for python) in IMAGE_INSTALL: http://pastie.org/1860360 May 03 13:47:30 mrAlmond, probably qtconfig (the only part of -common) is not built for qt4-embedded, thus the package is empty, does not get packed, and voila ... May 03 13:49:06 you cold check that in tmp/work//qt4-embedded_xxx/packages-split May 03 13:49:37 s/cold/could/ ;) May 03 13:49:51 this is my config http://pastie.org/1860372 May 03 13:50:01 mlip : :-D May 03 13:51:13 ls - l $OETOP/tmp/deploy//ipk/ | grep qt4-embedded-common, and see if it May 03 13:51:15 is there May 03 13:52:49 ka6sox-work: did you get any reply from GH people? May 03 13:56:37 mlip : There is no qt4-embedded-common* May 03 13:56:52 I thought it was the basic package of qt4-embedded May 03 13:57:52 mrAlmond, iff you are you using old org.openembedded.dev repository:http://cgit.openembedded.net/cgit.cgi/openembedded/tree/recipes/qt4/qt4.inc May 03 13:58:11 I'm using meta-openembedded May 03 14:00:31 mrAlmond, http://cgit.openembedded.net/cgit.cgi/meta-openembedded/tree/meta-oe/recipes-qt/qt4/qt4.inc , FILES_${QT_BASE_NAME}-xxx specify what's in their packages, look for what you need May 03 14:00:48 mlip : tnx! May 03 14:02:28 mrAlmond, yw May 03 14:18:24 I'm trying to build release-2011.03 and it's failing on attempting to download fakeroot_1.14.5. I've double checked the head of master and I don't see any mods I'm missing... Any ideas? May 03 14:19:26 thhp: create own recipes for fakeroot-1.15.1 May 03 14:19:50 or google for the old sources May 03 14:20:21 vanner_, kergoth : thanks May 03 14:20:30 so it's a known problem? May 03 14:20:43 ? May 03 14:20:49 we download from upstream May 03 14:20:55 upstreams sometimes remove old sources May 03 14:20:57 it happens May 03 14:21:09 doesn't imply that we have any knowledge about this particular one May 03 14:21:21 (though perhaps someone does, it's not necessary) May 03 14:22:00 understood May 03 14:22:53 I had a similar issue with ncurses which /is/ fixed in a later oe tree to use the oe mirror rather than upstream -- but in this case even the mirror failed, which is why I was confused May 03 14:54:04 thhp: What DISTRO are you using? Angstrom uses the angstrom mirror which has the sources May 03 14:54:15 No one else has stepped up and said they wish to maintain some other DISTRO in 2011.03 May 03 14:56:09 there are distros other than Angstrom :-D May 03 14:57:48 Tartarus: I have been defining my own distro based on "minimal" in order to use a third-party toolchain May 03 14:58:07 I didn't realise different distros maintain different mirrors, that's useful information May 03 14:59:27 the angstrom mirrors get fed regularly from real builds May 03 15:00:34 thhp: Then you might want to make your distro leverage the angstrom-mirrors class May 03 15:06:16 Tartarus, did you see koen's pull request? May 03 15:06:25 Tartarus: I'll have a look at that, many thanks May 03 15:06:44 #16? yes, giving it about a day, so shortly now I'll do the pull May 03 15:07:21 thanks May 03 15:09:59 03Florian Boor  07org.openembedded.dev * r0fd385ef55 10openembedded.git/recipes/ushare/ushare_1.1a.bb: May 03 15:09:59 ushare: We need lsb-base in the filesystem. May 03 15:09:59 ushare does not start up properly without scripts from lsb-base. May 03 15:09:59 Tested in µCross May 03 15:09:59 Signed-off-by: Florian Boor May 03 15:10:02 03Florian Boor  07org.openembedded.dev * r419cb686b0 10openembedded.git/recipes/lsb/lsb-base_3.2.bb: May 03 15:10:02 lsb-base: Package for LSB init-functions script May 03 15:10:02 Signed-off-by: Florian Boor May 03 15:10:09 03Florian Boor  07org.openembedded.dev * r95bf5692c5 10openembedded.git/recipes/ubootchart/ubootchart_svn.bb: May 03 15:10:09 ubootchart: Add recipe for SVN version May 03 15:10:09 A small and easy to use boot process profiler generating bootchart compatible output. May 03 15:10:09 From µCross, tested on armv5te platforms and other. May 03 15:10:09 Signed-off-by: Florian Boor May 03 15:10:10 03Florian Boor  07org.openembedded.dev * r35393002bb 10openembedded.git/recipes/gnome/zenity_2.28.0.bb: May 03 15:10:11 zenity: add `gnome-doc-utils` to `DEPENDS` May 03 15:10:11 Signed-off-by: Florian Boor May 03 15:10:12 03Florian Boor  07org.openembedded.dev * r4de1e9814a 10openembedded.git/recipes/gpe-lights/gpe-lights_0.13.bb: May 03 15:10:12 gpe-lights: fix license May 03 15:10:13 Signed-off-by: Florian Boor May 03 15:10:22 03Florian Boor  07org.openembedded.dev * r839711d6ed 10openembedded.git/recipes/zbar/zbar_0.10.bb: May 03 15:10:22 zbar: Add version 0.10 May 03 15:10:22 Signed-off-by: Florian Boor May 03 15:10:25 03Florian Boor  07org.openembedded.dev * rb3d5cd8306 10openembedded.git/recipes/proftpd/ (files/proftpd-install-nostrip.patch proftpd_1.3.3c.bb): May 03 15:10:25 proftpd: Add version 1.3.3c May 03 15:10:25 Update in order to get rid of several vulnerabilities of previous releases. May 03 15:10:25 Signed-off-by: Florian Boor May 03 15:14:42 Hum amazing... pushing a branch with a typo creates a new one instead of an error. May 03 15:15:11 florian: yup, thats what git does May 03 15:15:34 Amazing :-) May 03 15:16:27 one mans typo is another mans new branch :) May 03 15:18:00 Hum... I see what Paul means with "style" issues. The proftpd recipe is ugly... I guess I should check why it has to look that evil. May 03 15:59:01 otavio, just waking up to check. May 03 16:49:55 Hi :-) May 03 16:49:59 ka6sox-away: oh ok nice May 03 17:32:13 Is someone having problem compiling udev in meta-oe? May 03 17:32:14 | i586-poky-linux-libtool: link: cannot find the library `/home/otavio/hacking/el/tmp/sysroots/ossystems-x86/usr/lib/libusb-1.0.la' or unhandled argument `=/usr/lib/libusb-1.0.la' May 03 17:32:17 | make[2]: *** [extras/hid2hci/hid2hci] Error 1 May 03 17:32:20 | make[2]: *** Waiting for unfinished jobs.... May 03 17:34:36 with teh "=/usr/lib/..." it looks like you aren't using the stock compiler, linker -- or your libtool is broken May 03 17:34:48 The linker needs to understand the May 03 17:35:17 Libtool as well needs to know it, I thought it did May 03 17:35:33 (for those not familiar, the = syntax says use this library from within the sysroot) May 03 17:37:40 fray: in any case it seems as a bug in meta-oe since I am using the tools from it May 03 17:39:35 yes.. it does May 03 17:39:56 fray: yes and building udev from oe-core also fails with same error May 03 17:40:29 sounds like it's the linker or libtool then.. either of them been updated recently? May 03 17:40:49 fray: I started using oe-core two days ago ... no idea May 03 17:42:27 doing a full rebuild with empty sstate May 03 17:42:55 yup, sounds like an issue.. I havn't done a build since last week... May 03 17:43:28 fray: will let you know in a minute May 03 17:43:43 otavio: could always test a build without meta-oe as a sanity check :) May 03 17:45:17 kergoth: it would be nice if bitbake could output the layer it is grabbing the recipe May 03 17:45:33 kergoth: in any case i am doing a full rebuild to check May 03 17:45:48 otavio: i was thinking about another command for bitbake-layers, to summarize what's coming from where May 03 18:00:19 ka6sox-work: re SSD yes I am not using this atm May 03 18:00:26 ka6sox-work: so please go ahead May 03 18:38:31 03Michael 'Mickey' Lauer  07org.openembedded.dev * r3223f375c0 10openembedded.git/recipes/python/ (8 files in 2 dirs): May 03 18:38:31 PyQt: introduce 4.8.3; remove obsolete 4.4.3 May 03 18:38:31 This version of PyQt builds against Qt 4.7.x; please don't remove May 03 18:38:31 this version unless you also bring in a PyQt version that works. May 03 18:38:37 03Michael 'Mickey' Lauer  07org.openembedded.dev * r36b9a8ea02 10openembedded.git/recipes/python/python-sip.inc: May 03 18:38:37 python-sip: catch up with upstream changes May 03 18:38:37 upstream did change a lot, and with the next major version May 03 18:38:37 of sip they will change everything again. Take care when May 03 18:38:37 updating this recipe. May 03 18:38:53 03Michael 'Mickey' Lauer  07org.openembedded.dev * r6e4f126a92 10openembedded.git/recipes/sip/ (sip-native_4.12.bb sip-native_4.7.9.bb): May 03 18:38:53 sip-native: 4.12 -> 4.12.1; also remove version 4.7.9, which no longer is relevant May 03 18:38:53 sip is mereley used for PyQt bindings and the corresponding Qt version (4.4.x) May 03 18:38:53 has been removed many months ago. May 03 18:39:02 03Michael 'Mickey' Lauer  07org.openembedded.dev * r86edd16141 10openembedded.git/recipes/python/ (python-sip_4.12.1.bb python-sip_4.7.9.bb python-sip_4.8.2.bb): python-sip: 4.8.2 -> 4.12.1 May 03 18:58:08 hi. i'm trying to get learn openembedded. i've built an image for the nokia n900 and now i'm trying to figure out how i can try it out. any hints or recommendations on good howtos on this? May 03 18:59:04 re May 03 18:59:41 not offhand, but the flasher application for the n900 should be a good start May 03 19:10:27 the flasher does help, yes. i'm currently more interested in how to make the image May 03 19:11:11 bitbake didn't create a jffs2 image even though the nokia900 machine configuration says jffs2 is a target filesystem May 03 19:11:56 does your local.conf override this? May 03 19:12:14 good question. let's see May 03 19:12:36 IMAGE_FSTYPES += "tar.bz2" May 03 19:12:43 but i guess that's not an override May 03 19:12:48 just appending? May 03 19:14:02 indeed May 03 19:15:52 i don't see any fstype overrides in the image recipe either.. May 03 19:16:21 sampo_v2, hi May 03 19:16:28 i guess i can just make an ext2 partition an untar the .tar.bz2 image and the modules onto it May 03 19:16:32 hi May 03 19:16:50 I've an n900 too and I did some work for getting it in an usable state into oe May 03 19:17:02 what do you want to do with the n900 exactly? May 03 19:17:17 Note that SHR(an oe disrto) supports it well May 03 19:17:33 actually, i'm just using it as a testbed for the work that will be put on a beagle May 03 19:17:39 ah ok May 03 19:17:47 look in the machine config May 03 19:17:54 you have ubifs things May 03 19:18:07 including a comment for the right cmdline argument May 03 19:18:10 no ubifs for the nokia900 machine conf May 03 19:18:11 for the kernel May 03 19:18:22 machine conf at: May 03 19:18:43 conf/machine/nokia900.conf May 03 19:18:47 yes May 03 19:18:56 oh.. IMAGE_FSTYPES ?= "jffs2 tar.gz" May 03 19:18:57 # Add "rootfstype=ubifs ubi.mtd=5,2048 root=ubi0:nokia900-rootfs" May 03 19:18:58 # to the kernel cmdline to boot an ubifs image flashed on /dev/mtd5 May 03 19:19:17 i guess that means that because i have a IMAGE_FSTYPES in the local config, the jffs2 is not added May 03 19:19:25 s/added/put in/ May 03 19:19:40 I never tested jffs2, only ubifs on NAND May 03 19:19:50 ok May 03 19:20:24 then override IMAGE_FSTYPES in local.conf May 03 19:20:28 do you recommend using a microSD card or some other method? May 03 19:20:29 or += it May 03 19:20:34 yes of course May 03 19:20:42 ubifs would overwrite maemo May 03 19:21:00 :) May 03 19:21:32 so are those parameters supposed to be passed to flasher when booting the kernel? May 03 19:21:56 if you don't want to mess up maemo do that: May 03 19:22:03 *install the chainloaded meego uboot May 03 19:22:23 *use it to boot your image and also maemo May 03 19:22:26 when finished May 03 19:22:43 dowload the maemo Fiasco kernel image May 03 19:22:46 and flash it back May 03 19:23:02 or better May 03 19:23:14 since you already use the proprietary flasher and not 0xFFFF May 03 19:23:23 you can boot the combined uboot image May 03 19:23:28 without flashing it May 03 19:25:56 "combined uboot image", do you mean the "uImage-2.6.35-r58-nokia900.bin" file created by bitbake? May 03 19:26:37 no May 03 19:26:55 I mean meego's uboot + maemo's kernel combined in one file May 03 19:26:59 with the python script May 03 19:27:27 http://al.robotfuzz.com/~al/maemo/u-boot/ May 03 19:29:10 sampo_v2, note that we also have meego 37 kenrel in oe.dev May 03 19:29:27 that has proper power management May 03 19:29:30 umm. i'm not feeling very confident about messing this much with the n900 without understanding all of the steps May 03 19:29:40 (DVFS,CPUFREQ,SmartReflex etc...) May 03 19:29:51 sampo_v2, basically I can explain May 03 19:29:59 what you will do is the following May 03 19:30:08 1)wget uboot May 03 19:30:16 2)assemble uboot with maemo's kenrel image May 03 19:30:29 3)boot uboot trough flasher *without flahing it* May 03 19:30:50 4)from uboot boot the kernel from the microsd and your image May 03 19:31:04 so it won't touch the phone NAND at all May 03 19:31:08 if done correclty May 03 19:31:42 you don't even need to combine it tough May 03 19:33:23 someone should try to run kexecboot on n900 :) May 03 19:34:05 Jay7, we have a chainloaded uboot, what would be the advantages of kexecboot over it? May 03 19:34:08 * GNUtoo|laptop listens May 03 19:34:24 because indeed it would be great if it has a lot of advantages May 03 19:34:29 easy to use GUI :) May 03 19:34:49 mrmoku, ^^^ May 03 19:35:00 what do you think? May 03 19:35:09 maybe we switch to kexecboot? May 03 19:35:18 altough meego already uses uboot May 03 19:35:23 so I wonder.... May 03 19:35:51 Jay7, what's the footprint of kexecboot? May 03 19:36:04 I have switch_root and switch_root + losetup support in my todo May 03 19:36:22 then you may boot with GUI but w/o kexec May 03 19:36:29 GNUtoo|laptop: thanks for the explanation. my problem is just that most of the documentation on embedded stuff is deeply technical (too deep for a newcomer to learn what's what) and then there are high level tutorials which also fail to tell me why a step is done and what it does exactly. i'd be much more comfortable if i learned the why and not just how May 03 19:36:31 I know switch_root(the most dangerous command ever) May 03 19:37:06 sampo_v2, the why of what? May 03 19:37:07 GNUtoo|laptop: current zaurus kernels + kexecboot initramfs are about 800Kb May 03 19:37:16 we need to be under 2M May 03 19:37:17 with lzma compression May 03 19:37:21 with maemo + SHR kernel May 03 19:37:28 we need to be under 1.2Mb on zauruses May 03 19:37:34 GNUtoo|laptop: what does that mean for booting then? kexecboot from within maemo? May 03 19:37:35 i'd expect i could just make a root filesystem on an sd card, then somehow get the device to boot the kernel with the parameters so the kernel chooses the correct / May 03 19:37:43 mrmoku, no May 03 19:37:47 creating the root fs is easy May 03 19:37:49 I think we flash a kexecboot May 03 19:37:58 that sounds indeed nice May 03 19:38:01 mrmoku, instead of uboot May 03 19:38:01 mrmoku: flash kexecboot as 'primary' kernel into nand May 03 19:38:09 indeed May 03 19:38:12 and from what i've gathered the nokia flasher uploads the kernel to the device May 03 19:38:18 Jay7: ok May 03 19:38:28 then boot from any other media with kexec (right now), switch_root [+ losetup] (later) May 03 19:38:33 GNUtoo|laptop: I fear that won't work to boot maemo though May 03 19:38:43 ahh.. .hmm May 03 19:38:49 i don't understand why i would need or want anything more complex at this point (like u-boot, etc.) May 03 19:39:10 * mrmoku thinks he is confusing kexec and kexecboot probably May 03 19:39:14 Jay7, mrmoku basically the idea is to put maemo kenrel on initramfs and meego/SHR kernel on microsd May 03 19:39:33 GNUtoo|laptop: np, we are booting zauruses in that way May 03 19:39:34 mrmoku, kexecboot is a kexec-based bootloader: May 03 19:39:48 only show-stopper is ubifs absence in kexecboot right now :( May 03 19:40:04 mrmoku, you boot from NAND into the kexec-boot-kenrel that acts like a bootloader and that has an initramfs May 03 19:40:15 mrmoku: http://kexecboot.org/about :) May 03 19:40:29 and /screenshots May 03 19:40:33 Jay7, what do you mean by ubifs absence? it can't load a kernel that is on ubifs? May 03 19:40:48 GNUtoo|laptop: it can't detect ubi right now.. May 03 19:41:09 it's possible with some hacks May 03 19:41:12 JaMa, does it support ubi kenrel arguments? May 03 19:41:16 like for booting maemo May 03 19:41:17 oops May 03 19:41:18 sorry May 03 19:41:20 Jay7, May 03 19:41:21 Jay7: yeah, already reading :) May 03 19:41:30 GNUtoo|laptop: kernel is ready to this May 03 19:41:32 I mean GUI May 03 19:41:36 ok May 03 19:41:36 nice May 03 19:41:37 kexecboot itself May 03 19:42:09 so basically we boot the maemo kenrel that reside on initramfs, and pass ubifs root= arguments and it'll work May 03 19:42:12 * Jay7 should update screenshots btw.. May 03 19:43:04 GNUtoo|laptop: something like.. but I'll prefer to have ubifs detection code in kexecboot.. May 03 19:43:16 Jay7, does kexecboot support some kind of network kenrel, like NFS, or similar? May 03 19:43:38 GNUtoo|laptop, Jay7: if we get it to work I'm all for it... looks very nice indeed May 03 19:43:39 that would ease development a lot May 03 19:43:57 GNUtoo|laptop: nfs is impossible with kexec if I understand it right.. so only with switch_root May 03 19:44:17 Jay7, I mean something like in android where you send the kernel trough USB May 03 19:44:26 in order to boot without flasing May 03 19:44:30 *flashing May 03 19:44:46 GNUtoo|laptop: ah.. right now - no May 03 19:44:50 ok May 03 19:45:41 so I think we should go kexecboot anyway May 03 19:45:42 so... May 03 19:45:49 how do we make it work within oe May 03 19:45:58 just try before May 03 19:46:00 we need a device specific kenrel defconfig for kexecboot I guess May 03 19:46:14 you may compile kexecboot as a stand-alone binary May 03 19:46:16 that needs to be small enough May 03 19:46:22 and run from some rootfs May 03 19:46:46 just to check FB UI to work May 03 19:47:06 we have no N900 around, so it's not tested before May 03 19:47:14 no problem May 03 19:47:27 but I wonder where is the kexecboot kernel config stored in oe May 03 19:47:36 linux-kexecboot* May 03 19:47:40 ah ok May 03 19:47:42 thanks a lot May 03 19:48:04 ant is our kernel guy :) May 03 19:48:11 ok May 03 19:48:15 kernel/oe infra May 03 19:48:53 did you test on omap? May 03 19:48:56 like beagle? May 03 19:49:01 I see no beagle defconfig May 03 19:49:12 GNUtoo|laptop: Koen have screenshot from beagle May 03 19:49:12 only zaurusses May 03 19:49:15 ok May 03 19:50:13 is there an howto or explanation on how to get oe images to boot on a beagle? May 03 19:52:01 mrmoku, JaMa seem to have ported it to SHR already....strange May 03 19:52:46 GNUtoo|laptop: hm.. seems there was support of 2.6.32 for beagle May 03 19:52:55 check git logs May 03 19:53:06 http://lists.linuxtogo.org/pipermail/openembedded-commits/2010-January/040433.html May 03 19:53:08 ok May 03 19:53:35 http://www.flickr.com/photos/koenkooi/3761596514/ May 03 19:53:52 screenshot of old (0.4) kexecboot ui on beagle May 03 19:54:04 well.. not screenshot but phote May 03 19:54:09 *photo May 03 20:14:25 I can't get perl-native to compile. Is this a known issue? May 03 20:17:05 khem, can you live in 107GB up from 89GB? May 03 20:18:10 ka6sox-away: surely May 03 20:18:22 ka6sox-work: that should be enough to start with May 03 21:39:40 hi Jay7 I've an issue with kexecboot: the maemo kernel is somehitng like 1.7M and I didn't try lzma -9 on zImage but I think it won't reduce it enough May 03 21:40:10 so maybe ubifs support would be great to add May 03 21:40:16 is it that hard to add May 03 21:40:17 ? May 03 21:40:25 even if it's a hack or something like that May 03 21:40:56 03Chris Larson  07master * r51872fce3f 10openembedded.git/recipes/gstreamer/gst-plugins-package.inc: May 03 21:40:56 gst-plugins-package: fix with package_dbg May 03 21:40:56 Signed-off-by: Chris Larson May 03 21:41:42 GNUtoo|laptop: to be able to kexec things you don't need build full kernel May 03 21:41:56 we are using stripped defconfigs for zauruses May 03 21:42:01 about ubifs May 03 21:42:34 GNUtoo|laptop: it should be possible to call ubiattach and then just force using ubifs w/o using fs detection routing May 03 21:42:38 *routine May 03 21:42:55 as ugly hack :) May 03 21:43:01 ok thanks a lot May 03 21:43:04 I'll build it now May 03 21:43:10 I just need a defconfig right? May 03 21:43:14 or do I need something else? May 03 21:43:27 yes, place defconfig alike it's done for Zauruses May 03 21:43:35 then build linux-kexecboot May 03 21:43:59 or just build kexecboot to test as standalone binary May 03 21:44:35 ok May 03 21:44:56 btw, you should add n900 to kexecboot.inc May 03 21:45:13 --enable-uimage is useful for n900 :) May 03 21:45:19 it is disabled by default May 03 21:45:58 ok May 03 21:46:00 thanks a lot May 03 21:46:08 np, feel free to ask May 03 21:46:18 we may talk on #kexecboot channel May 03 22:04:10 ERROR: Unable to unpack '/home/gnutoo/embedded/oe/org.openembedded.dev/recipes/linux/linux-kexecboot-2.6.38/././logo_linux_clut224.ppm.bz2 May 03 22:04:19 and I set logo in the machine config like that: May 03 22:04:57 LOGO_SIZE = "vga" May 03 22:05:22 moment May 03 22:06:33 GNUtoo|laptop: iirc, LOGO_SIZE should be enough May 03 22:06:48 hello, Andrea May 03 22:07:16 GNUtoo|laptop: check machine/ben-nanonote.conf for #ScreenInfo if not May 03 22:07:50 and, btw, for require at bottom May 03 22:08:20 GNUtoo|laptop: ant__ may say more on how to add kexecboot support to machine in OE :) May 03 22:08:30 ok May 03 22:08:53 ant__: we are trying to test kexecboot on nokia n900 :) May 03 22:10:27 it still put the logo as . May 03 22:15:44 ahhh May 03 22:15:54 I was building for om-gta02 May 03 22:16:06 sorry May 03 22:16:16 I forget to switch to nokia900 May 03 22:20:53 * XorA|gone wonders if we can get any troll poison for the next OEDEM May 03 22:21:03 | lzma: invalid option -- 'e' May 03 22:21:17 | + lzma -c -e -9 May 03 22:23:49 I'll fix in local.conf May 03 22:23:58 s/fix/workarround/ May 03 22:24:01 GNUtoo|laptop: we have it overriden in machine confs iirc May 03 22:24:07 ah ok May 03 22:32:00 yes, otavio's patch still needing one more ack May 03 22:43:53 black screen May 03 22:44:03 illuminated black screen May 03 22:44:07 I bet it's not normal May 03 23:30:48 03Petr Štetiar  07org.openembedded.dev * r36dfc09848 10openembedded.git/recipes/modutils/modutils-initscripts.bb: (log message trimmed) May 03 23:30:48 modutils-initscripts: fix wrong order of module loading happening in udev May 03 23:30:48 In the current state, udev init script is loaded as 3rd and modutils.sh as May 03 23:30:48 20th, so udevadm called in udev init script loads all kernel modules in some May 03 23:30:48 unpredictable order, not honoring in any way specified order of modules in May 03 23:30:49 /etc/modules. This causes some troubles mainly in the first boot. So to fix May 03 23:30:50 this we now move loading of the kernel modules just before we exec udev init May 03 23:33:31 otavio, ping? May 03 23:37:11 hmmm...now I need to figure out how to limit our # of commit messages to 5 in a row before it takes a breather. May 03 23:41:29 03Andrea Adami  07org.openembedded.dev * r3238d0e004 10openembedded.git/recipes/udev/udev_162.bb: May 03 23:41:29 udev: modify udev_162.bb INITSCRIPTS_PARAMS, following oe-core May 03 23:41:29 * initscripts expects udev to start at S04, after S03sysfs.sh May 03 23:41:29 * udev_cache is moved at S36 after S35mountall.sh May 03 23:41:29 Signed-off-by: Andrea Adami May 03 23:41:40 03Andrea Adami  07org.openembedded.dev * r974dfb8777 10openembedded.git/recipes/udev/udev_165.bb: May 03 23:41:40 udev: modify udev_165.bb INITSCRIPTS_PARAMS, following oe-core May 03 23:41:40 * initscripts expects udev to start at S04, after S03sysfs.sh May 03 23:41:40 * udev_cache is moved at S36 after S35mountall.sh May 03 23:41:41 Signed-off-by: Andrea Adami May 03 23:44:34 03Andrea Adami  07org.openembedded.dev * r100f383771 10openembedded.git/recipes/udev/ (udev_162.bb udev_165.bb): May 03 23:44:35 udev: bump PR of 162 and 165 recipes following initscripts changes May 03 23:44:35 * only those recent versions have been updated May 03 23:44:35 * versions <1.6 are actually not actively supported anymore May 03 23:44:35 Signed-off-by: Andrea Adami May 03 23:45:15 worth a clean rebuild now... May 03 23:46:02 hello can anyone give some hints on compiling? May 04 00:00:21 khem, for some reason (unk) garnet is being a pill. May 04 00:01:02 I havent touched it in days May 04 00:01:43 no, this is something else...but I keep cranking up the log level and still no telling May 04 00:01:58 might have to rebuild the initramfs May 04 00:02:08 khem, any reason this can't be 64bit? May 04 00:20:11 anyone using python 2.7? I see the recipes set python-2.6 preference to -26 and python-2.7 preference to -27 - that makes python 2.6 preferred - not clear why? May 04 01:19:09 tharvey: silly math :) May 04 01:19:23 1 is preferred over 0 over -1 over ... over -26 over -27 May 04 01:19:31 So if nothing has 1 (which is implicit) May 04 01:20:12 right, so thats why python 2.6 is preferred right? (because -26 > -27) May 04 01:20:42 not clear why someone would have given them those preferences... would have expected them to be positive values or not specified at all so that the newer python 2.7 was preferred? May 04 01:21:13 Well, imho jumping to 2.7 over 2.6 needs some testing May 04 01:21:17 since not everything works May 04 01:21:32 If it was intentional silliness or not, I don't know May 04 01:22:49 hi , i'm having some problem , i want to develop some application that will run Linux first and then Linux runs some my custom code and boot windows back , is it possible ? if yes , how ? May 04 01:22:54 maybe u know this problem May 04 01:27:14 JDuke128, you could do that by changing the bootloader to boot the other OS from your application May 04 01:27:35 should be easy enough for e.g. uboot May 04 01:33:12 Tartarus, thats what I was wondering... if the recipes simply had not been tested yet. I still wonder about the preference setting though, seems like if you wanted to prefer 2.6 over 2.7 you would use PREFERRED_VERSION or make DEFAULT_PREFERENCE=-1 in the 'untested' python 2.7 May 04 01:35:05 yes May 04 01:36:18 thx May 04 01:37:42 Tartarus, would you be able to run a buildbot slave @ your builder site? May 04 01:38:01 long term? not really May 04 01:38:25 i'd have to tie the jenkins dynamic allocation to buildbot and vice versa May 04 01:43:52 okay, nm then... May 04 01:44:21 I'll just continue with working on a buildstats class. May 04 02:00:47 Hello - question. I want to reflash the root filesystem on my device with mkfs.jffs and tar x. I've copied the tools I need into a tmpfs, now I want to switch_root into that tmpfs, but it doesn't work. May 04 02:00:58 How do I switch to the new root and unmount the old one? May 04 02:08:20 Ok, I think I got it, but now I'm seeing: mkfs.jffs2: mkfs.jffs2://proc/10/exe: No such file or directory **** ENDING LOGGING AT Wed May 04 02:59:58 2011