**** BEGIN LOGGING AT Thu May 20 03:00:00 2010 May 20 03:38:23 * kergoth sighs May 20 03:38:36 * kergoth is going to have to modify bitbake to enable analysis and hashing of methodpool functions May 20 03:39:02 hmmmm May 20 03:41:04 wiat.. May 20 03:41:06 wait, even.. May 20 03:41:25 i think i could pickle.dumps() the function to at least hash the methods in the pool.. May 20 03:41:39 can't analyze them to determine the variables they reference or anything, so that'd have to be explicit, but.. May 20 03:41:42 * kergoth ponders **** BEGIN LOGGING AT Thu May 20 03:51:34 2010 May 20 04:58:30 Hi, ALL May 20 04:59:24 I wrote a bitbake recipe to compile my own program, however when issuing build the linker refer to the local /usr/local/lib May 20 05:00:06 and does not search for a path in an OE tree. Any idea on what might be the problem? May 20 05:00:35 if it's your program, you should know what the makefiles are doing May 20 05:00:42 so you're far more likely to have a clue about that than we will May 20 05:00:53 make sure the build is obeying the bitbake CC, CFLAGS, and LDFLAGS variables. May 20 05:02:27 kergoth: locally I created a KDevelop project, compile and run application May 20 05:02:39 okay? May 20 05:02:47 that doesn't make us somehow know what your buildsystem is May 20 05:03:00 In the recipe I just point the SRC_URI variable to the main path May 20 05:03:01 i assume kdevelop generates makefiles. read them May 20 05:03:24 again, no one is going to know how your app is built better than you will May 20 05:04:45 its autoconf/makefile project? May 20 05:19:12 kergoth: that's the problem, its the only thing that is different from other path in the compile/link command May 20 06:41:19 what is the purpose of the -isystem option? Where it should point to? May 20 06:43:30 man gcc ? May 20 06:43:40 goodmorning May 20 07:16:49 Quick question: Is GuruPlug supported by OE? If I set the machine to "SheevaPlug", will the builds work with GuruPlug? May 20 07:29:46 gmorning. i experience some problems with two custom machines and minimal distro config/glibc. somethime the networt does not start up. a reboot fixes the issue. any ideas? May 20 07:30:27 good morning May 20 07:31:07 doragasu: GuruPlug and SheevaPlug are different, so the answer is no May 20 07:34:32 hi May 20 07:41:31 hi ericben May 20 07:42:49 hi mckoan May 20 07:43:50 03Khem Raj  07org.openembedded.dev * r236f580a53 10openembedded.git/recipes/eglibc/ (eglibc-initial_2.12.bb eglibc_2.12.bb eglibc_svn.bb): May 20 07:43:51 eglibc-2.12: Add new recipes for 2.12 May 20 07:43:51 * Update svn recipe to latest revision on trunk. May 20 07:43:51 Signed-off-by: Khem Raj May 20 07:43:51 nik0n: I guess a bad driver in the kernel May 20 07:44:00 03Khem Raj  07org.openembedded.dev * r89019eab67 10openembedded.git/recipes/uclibc/ (3 files in 3 dirs): May 20 07:44:00 uclibc-git: Update machine config for SH3 and SH4 May 20 07:44:00 * Bump the SRCREV to latest on git master May 20 07:44:00 Signed-off-by: Khem Raj May 20 07:45:29 mckoan: yesterday we were discussing of the way to generate things like this : http://opensource.palm.com/1.4.1.1/index.html for an image, do you have such a thing in KaeilOS do avoid duplicate work ? May 20 07:47:01 ericben: "things like this" what? the table or the patches? May 20 07:47:08 mckoan: but why does it work sometimes? the driver is igbe so i think its not "bad" May 20 07:47:32 nik0n: so check the hardware ;-) May 20 07:48:54 mckoan: a way toi generate this kind of table when building the image. Denix has started a script in Arago project which is a good starting base but maybe you had the same need in Kaeil (I didn't check) May 20 07:48:56 mckoan: tripple checked. netbase initscript yells: ifup interface xy already configured May 20 07:48:59 strange... May 20 07:50:09 ericben: sorry I haven't anything like that May 20 07:50:53 nik0n: hmmm I would need more info and a board on my desk ;-) May 20 07:54:52 morning May 20 08:00:30 hi hrw May 20 09:57:45 hi stefan_schmidt May 20 09:58:29 hi ant_work May 20 09:58:33 btw openmoko-dfu-util works just fine on Gentoo/x86 May 20 09:58:50 ant_work: thanks, goo to know May 20 09:59:00 ant_work: did you have success with the bt dongle? May 20 09:59:06 they use svn5484 May 20 09:59:18 ant_work:could you try the rc1? May 20 09:59:21 well, I see both my dongles DFU-enabled May 20 09:59:31 one is CSR other is Broadcom May 20 09:59:41 ant_work: but not the one of your friend? May 20 09:59:53 ah, I don't know what he's doing May 20 10:00:06 I sent him my binary (tested on ubuntu livecd) May 20 10:00:10 mine at home is ROM May 20 10:00:26 I got a BC4 Ext on CSR May 20 10:00:29 ;) May 20 10:00:34 ant_work: Where you able to get the firmware from the dongles with dfu-util? May 20 10:00:47 yeah, mine ist BC4 but ROM May 20 10:00:55 Need to check out for another one some time May 20 10:00:56 I did not yet try to upgrade/downgrade this one May 20 10:01:34 ant_work: Would be good to know if you are able to upload from the device. I don't ask people to upgrade with it :) May 20 10:02:23 I suppose the CSR one will do anything I ask him May 20 10:02:34 as pid/vid of vanilla CSR :) May 20 10:02:43 012/001 May 20 10:02:56 heh May 20 10:03:07 the broadcom is tough, there are no utils.. May 20 10:03:24 and itseems to expose the HID profile in a different way May 20 10:04:03 I hope today in 2010 any el-cheapo dongle would accept HID proxy at boot May 20 10:04:13 or how they call they version May 20 10:06:39 ant_work: still working on kexecboot? May 20 10:06:58 yes, should be finished in June May 20 10:07:31 Jay7 is sponsored by CELF May 20 10:07:31 ant_work: http://people.canonical.com/~lool/dmart-uds-m-notes/arm-m-soft-bootloader.txt are notes from ubuntu session about mukluk (Ubuntu kexec based bootloader) May 20 10:07:36 I know May 20 10:08:49 ant_work: maybe would be worth for kexecboot to check from time to time how it goes and take interesting features May 20 10:08:53 basically JAy7 just need to add some forgotten things (initrd support) and finish the debug/ui May 20 10:09:24 someone should take care of the 'skin' too May 20 10:10:24 hrw: https://wiki.ubuntu.com/Specs/ARMSoftbootLoader May 20 10:10:45 then Michael Casadevall started QI May 20 10:10:59 too many bootloader around now ;) May 20 10:12:07 hrw: in the meanwhile we have LZMA compression so *possibly* busybox could fit in the initramfs on Zaurus too May 20 10:12:42 ant_work: that spec is after notes. I did not looked for it yet May 20 10:13:18 having shell and utils would make all easier...we already have such boot-images in OE May 20 10:15:05 yep May 20 10:15:56 in the last month there have been big fixes for kexec-tools (kexec and kdump) and for klibc May 20 10:16:16 for arm and now also for powerpc May 20 10:16:36 we'll have to synch OE asap May 20 10:18:06 stefan_schmidt: should I try git or rc1 sources? May 20 10:18:25 ant_work: rc1 please May 20 10:18:30 ok, I'll try to upload May 20 10:18:38 ant_work: great, thanks May 20 10:19:20 thank you again May 20 10:19:36 that guy is a bit problematic but he's a genius wrty OSX May 20 10:20:14 he's putting together some strange EFI Bioses/bootloaders for vanilla Hackintosh May 20 10:26:56 ant_work: yeah, was a strange communication with him. Genius problem perhaps... ;) May 20 10:34:24 hi woglinde May 20 10:35:45 jo stefan May 20 11:00:34 hi May 20 11:03:42 I have a question, is there a way to bypass the do_fetch action for all recipes ? Let's say I want to rebuild a kernel I made changes to in the work folder and I don't want these changes to be overwritten when I rebuild the complete image. Is there a quick/dirty way to do that or do I have to modify all the recipes I want to change to apply my changes as a patch before the do_compile ? May 20 11:06:01 chli for quicktesting May 20 11:06:18 run bitbake -c do_compile linux-bla May 20 11:06:19 ups May 20 11:06:20 -f May 20 11:06:28 -f == force May 20 11:06:33 so the other steps arent involved May 20 11:07:09 ok, but that would just rebuild that package, if I want to regenerate an entire image that's not an option. May 20 11:07:14 03Koen Kooi  07org.openembedded.dev * r7425e92b1c 10openembedded.git/classes/kernel.bbclass: May 20 11:07:15 kernel bbclass: switch to u-boot-mkimage-native May 20 11:07:15 This is a long demanded change. The u-boot-mkimage is up to date to OE standards, while the openmoko one is not May 20 11:07:25 03Koen Kooi  07org.openembedded.dev * r0394970076 10openembedded.git/recipes/u-boot/ (2 files in 2 dirs): u-boot-mkimage: convert to new style staging and bbclassextend May 20 11:07:32 chli right May 20 11:08:00 hm May 20 11:08:04 let me think May 20 11:08:11 the method you told me is what I use today :) May 20 11:08:11 you could do it May 20 11:08:25 run bitbake -c clean image-foo May 20 11:08:31 bitbake image-foo May 20 11:08:43 after you rebuild all your recipes with changes May 20 11:08:46 hi gnutoo May 20 11:08:56 what could cause "qemu: uncaught target signal 11 (Segmentation fault) - core dumped" during locale generation? May 20 11:09:03 qemu build by oe May 20 11:09:13 host x86_64 May 20 11:09:23 vonami which distro? May 20 11:09:28 ubuntu May 20 11:09:41 10.0.4? May 20 11:09:44 ups 10.04 May 20 11:09:48 yep May 20 11:10:03 hm May 20 11:10:11 which target? May 20 11:10:16 i686 May 20 11:10:24 args May 20 11:10:34 woglinde I don't think that's what I want :) But thanks for helping. May 20 11:10:34 why the hell qemu told for this May 20 11:10:46 args May 20 11:10:52 s/told/called May 20 11:11:14 I thought maybe there is a way to globally override the do_fetch to be skipped to avoid overwritting any source code changes made in the work folder May 20 11:11:35 chli no May 20 11:12:14 woglinde: I didn't reset the default value in mmap_min_addr. that could be it May 20 11:12:28 vonami but bitbake should complain May 20 11:12:32 * CHli is sad May 20 11:12:33 CHli: you can override the SRC_URI with the sourcefolder May 20 11:12:40 hm May 20 11:12:42 and anyway qemu is doing it self in the latest version in oe May 20 11:13:00 woglinde: latest as in? May 20 11:13:17 koen: do you ack the patch to linux.inc wrt conflicting getline? just sed or some complex version check? May 20 11:13:21 oh/ May 20 11:13:31 Heinervdm, can you elaborate ? How would that work ? May 20 11:13:32 woglinde: pls ack you als zweite ;) May 20 11:13:34 ant koen isnt here May 20 11:14:08 yep wrong label May 20 11:14:37 ant hm May 20 11:14:46 seems I deleted the mail already May 20 11:14:51 2.6.2x is broken atm May 20 11:15:01 CHli: copy the work folder to some other place and then add SRC_URI-pn_RECIPE = "file:///path/to/folder" to your local conf May 20 11:15:03 np, I'll bump May 20 11:15:39 Heinervdm, that sounds very good, where can I get more info about that feature ? May 20 11:15:42 CHli: but then you have to do -c clean after every change in that folder, or bump the PR May 20 11:16:47 CHli: don't know if there is any information about this May 20 11:18:10 CHli: this is what we have at SHR: http://www.shr-project.org/trac/wiki/Building%20SHR#Makingchanges May 20 11:20:28 ok, then you have to explain me a little bit more :) Do I understand right if one line must be added in local.conf for every packages I edited the source code in the work folder ? May 20 11:20:51 oh thanks Heinervdm I will have a look. May 20 11:23:00 Thanks Heinervdm, that's pretty much what I wanted, exept I thought it was maybe possible to override globaly. But that will do it for the moment. May 20 11:23:45 re May 20 11:33:21 jo florian May 20 11:40:52 thx woglinde May 20 13:06:14 How to speed up the compilation process done by bitbake ? May 20 13:06:58 ananth_: 1) buy faster system May 20 13:07:04 2) use BBTHREADS May 20 13:07:08 3) use SSD May 20 13:07:21 4) parallel make May 20 13:07:43 buy faster machine May 20 13:07:57 regarding 2) BBTHREADS where to change in receipe ? May 20 13:08:27 recipes takes PARALLEL_MAKE May 20 13:08:29 he...build in tmpfs and some races will arise ;) May 20 13:09:21 In bitbake allows multiple recipe build simultaneously ? May 20 13:09:28 is possible ? May 20 13:09:38 yes, usually works too May 20 13:10:28 I tried to do that like this bitbake omap3-console-image in one terminal and in other terminal i started bitbake omap3-desktop-image May 20 13:10:53 Both started compiling gcc-cross-intermediate_4.3.3.bb and end up in error some *.o file missing May 20 13:11:00 ah, you mean building multiple image, not recipes... May 20 13:11:05 so what you suggest to avoid this error ? May 20 13:11:15 that's different story May 20 13:12:01 we were talking about how to parallelize the tasks of *one* bitbake command May 20 13:12:16 you cannt compile two images at the same time May 20 13:12:57 well, if youreally want you need two distinct toplevel dirs May 20 13:13:12 but why? May 20 13:13:17 oh! where to enable parallel make May 20 13:13:48 see build/local.conf May 20 13:14:17 parallel_make is in the recipes as said above May 20 13:15:52 PARALLEL_MAKE = "-j 4" will uncomment be enough ? May 20 13:16:19 yes May 20 13:16:33 oh! Thanks ant_work May 20 13:16:42 if one recipe has different needs then this setting will be overriden May 20 13:16:47 in teh recipe May 20 13:17:20 other buildsystems are a bit smarter..I've read about bsd, it maxes out the make tasks dinamically May 20 13:18:31 oh! May 20 13:20:46 playing with jobs and max-load May 20 13:23:15 In local.conf i came across DISTRO_TYPE "release" and "debug" (What changes will this make in the output binary) May 20 13:23:48 good question May 20 13:23:55 Just optimization flags in the compilation flags (or) some thing more than that ? May 20 13:24:15 I fear that (Angstrom) keeps the gdb debug options in release too May 20 13:24:30 long that I don't build release :) May 20 13:24:56 oh! ok May 20 13:25:11 ananth__: just grep for 'release' and 'debug' in openembedded May 20 13:25:19 iirc is a distro thing May 20 13:26:41 ok May 20 13:32:42 RP, http://www.electronicsweekly.com/Articles/2010/05/20/48680/mathworks-tunes-open-source-plans-through-eclipse.htm May 20 13:32:46 I build open-embeeded desktop image (enlightment as gui manager) will my qt application will use enlightment gui manager ? May 20 13:35:48 ananth__: what is your understanding of GUI manager? May 20 13:37:09 if you want you can mix libs and have QT/x11 images as it was on Zaurus long ago May 20 13:49:51 guys: how old your $HOME dir is? I found files from 2001 in mine May 20 13:51:24 hrw: I'm pretty sure I have older ones :) May 20 13:52:28 florian: 12714215 4 -rw-r--r-- 1 hrw hrw 299 gru 31 1979 ./devel/OH/klienci/ST/ST-poky/meta-st/packages/pkg-update/files/gconf.tar.ff May 20 13:52:51 but such ones does not count May 20 13:53:00 well, cmos mismatch is not accepted here May 20 13:53:03 :) May 20 13:57:00 simpad starts with 1936 May 20 13:57:32 nice pet May 20 14:02:20 jo kergoth May 20 14:03:14 hey May 20 14:03:43 kergoth thanks for fixing the 2 javastuff recipes May 20 14:05:13 woglinde: 36? May 20 14:05:40 found 2000y files - my master thesis May 20 14:06:29 hrw *g* May 20 14:07:24 29.09.2000 - gif pictures May 20 14:08:06 woglinde: heh, was testing my variable ref tracking stuff, it parses the shell to determine what other shell functions it calls, and the parser choked on those :) May 20 14:09:57 kergoth *g* May 20 14:12:21 void DoSomething(void) :D May 20 14:14:55 my next worry is OVERRIDES changes. there are tasks that modify OVERRIDES and run update_data, i.e. to collapse specific vars.. but I think those can all be replaced by explicit uses of the pkg specific variables, so think I'm going to try to remove all of those today May 20 14:15:40 hrm May 20 15:18:15 morning kergoth May 20 15:18:20 jo pb May 20 15:18:25 hi woglinde May 20 15:23:52 pb the caravn already arrived? May 20 15:23:57 ups caravan May 20 15:37:04 Hello. I'm trying to build boost with python for an armv7a on a 64-bit system. The problem is that python does a check in a header file to make sure that a LONG_BIT definition matches the compiler's length for a long. How might I try to fix this? For example (maybe this is on the wrong track, but,) is there a way to tell bitbake to only build for the TARGET arch and skip the HOST staging completely? May 20 15:38:09 emm386 hm intressting problem May 20 15:40:10 woglinde: not yet, probably another couple of weeks May 20 15:40:26 emm386 one solution would be to use qemu May 20 15:40:33 like locales building for glibc May 20 15:41:08 ouch. that means setting up an oe from scratch. May 20 15:41:25 (and a vm...) May 20 15:41:55 emm386 oh your question was not oe specific May 20 15:43:27 woglinde: I should have specified, I'm building using oe, and the bitbake script for boost. May 20 15:43:45 oh May 20 15:44:01 I meant not running all with qemu May 20 15:44:07 Good morning ALL May 20 15:44:08 only the python compiler test May 20 15:44:13 hi IgorK May 20 15:44:34 emm386 or perhaps you can fool the test May 20 15:44:49 In the make file do I refer to OE/temp/rootfs or to OE/temp/staging? May 20 15:45:12 IgorK which makefile? May 20 15:45:20 rootfs is out of scope May 20 15:45:32 its only used for fs-creation May 20 15:45:44 I am talking about makefile for my program. OE is failing me ;-) May 20 15:46:01 use autotools May 20 15:46:01 woglinde: looking into that now. the file is staging//usr/include/python2.6/pyport.h around line 675 May 20 15:46:38 yes, in the recipe it is inherited and the compile/link command is still wrong May 20 15:47:28 IgorK inherit autotools dont make your Makefile autools aware May 20 15:48:28 emm386 uh this test is from 2000 May 20 15:49:01 woglinde: I am using the sources from the projetc generated by KDevelop in x86 Gentoo Linux May 20 15:49:03 IgorK are you sure your program has all the autotools stuff? May 20 15:49:27 kdevelop is producing sane configure.ac and Makefile.am's? May 20 15:49:29 and it is compiled fine at x86 regular gcc May 20 15:49:40 this means nothing May 20 15:49:45 yes for x86 gcc May 20 15:49:49 tons of software compiles fine native May 20 15:49:55 and chokes at crosscompiling May 20 15:50:31 well its not choking it generates wrong makefile May 20 15:50:34 because the dev's wasnt so smart to take the 3 to 4 days to learn autotools or cmake May 20 15:51:28 is it possible to avoid oe to build its kernel package ? May 20 15:51:31 I guess its because everything is created for the x86, right? May 20 15:52:19 are you using KDevelop Automake manager? May 20 15:52:27 I just want build a rootfs without kernel/modules May 20 15:52:58 djdeath I fear not, you have to make an imagr hook which removes the stuff from rootfs May 20 15:53:05 ups image hook May 20 15:53:23 In KDevelop I just hit F8 and it runs configure and make May 20 15:53:39 djdeath look at /devel/arm/git/openembedded/recipes/images/beagleboard-demo-image.bb May 20 15:53:54 I didn't change anything in the project it self May 20 15:53:57 ROOTFS_POSTPROCESS_COMMAND May 20 15:54:26 I guess it's using it May 20 15:54:34 djdeath maybee there is a ROOTFS_POSTPROCESS_COMMAND which removes the kernel stuff May 20 15:54:45 igork hm yes May 20 15:54:58 maybee you can post the configure.ac/in May 20 15:55:28 http://www.kdevelop.org/mediawiki/index.php/FAQ#Where_is_the_Automake_manager_in_KDevelop.3F May 20 15:55:35 woglinde: I have seen that using an external toolchain you can avoid some packages May 20 15:55:58 djdeath hm only for compiling May 20 15:58:05 re hrw May 20 15:58:14 re May 20 15:58:18 checking xchat May 20 16:00:01 woglinde: configure.in posted: http://pastebin.com/BW8pQ9UC May 20 16:00:20 Im using it to add new files to the project... May 20 16:01:33 igork May 20 16:01:35 # May 20 16:01:35 CPPFLAGS="$CPPFLAGS -Wall -g -fexceptions" May 20 16:01:35 # May 20 16:01:37 CXXFLAGS="$CXXFLAGS -Wall -g -fexceptions" May 20 16:01:51 this should be placed in Makefile.ams May 20 16:03:43 woglinde: I don't have such file do you mean in Makefile.am? May 20 16:04:04 yes May 20 16:04:42 igork read http://www.gnu.org/software/automake/manual/automake.html#Flag-Variables-Ordering May 20 16:16:54 IgorK and I would factor out the wx-stuff into its own m3 file May 20 16:16:58 ups m4 file **** BEGIN LOGGING AT Thu May 20 16:18:43 2010 May 20 16:18:47 igork the cacao project is using autotools in a really good way May 20 16:18:51 have a look http://mips.complang.tuwien.ac.at/hg/cacao/file/8948a434c10d May 20 16:47:16 woglinde: this is the command that KDevelop run on brand new project to generate configure: http://pastebin.com/1MEmsrv2 May 20 16:47:45 I think I need some thing like this to run inside OE to generate proper configure script May 20 16:47:55 WANT_AUTOMAKE_1_6="1" May 20 16:48:00 thats stoneage May 20 16:48:20 sorry May 20 16:48:35 but dont know whats kdevelop is doing there May 20 16:48:40 thats OK, system didn't updated in a long time May 20 16:48:44 try to understand autotools May 20 16:48:50 and make the stuff your self May 20 16:48:57 autotools isnt that hard May 20 16:49:03 especially for your project May 20 17:00:27 hi all! May 20 17:06:16 jo zecke May 20 17:07:43 03Tom Rini  07org.openembedded.dev * rb4f2c5f863 10openembedded.git/conf/distro/include/sane-toolchain.inc: May 20 17:07:43 sane-toolchain.inc: Don't hold ppc405 back to gcc 4.4.1 May 20 17:07:43 Given the other powerpc issues that are now fixed in 4.4.4, this leads to May 20 17:07:43 ppc405 being broken. May 20 17:07:43 Signed-off-by: Tom Rini May 20 17:10:51 hm libtool is using legacy do_stage too May 20 17:14:55 is stable/2009 still recommended for angström builds? May 20 17:23:07 RP: ping May 20 17:55:26 Hi all May 20 17:55:46 I managed to build a console-image May 20 17:56:13 I would like to do my own 'image' copying console image and adding some more packages May 20 17:56:24 where is located console-image? May 20 17:56:34 is it a .bb recipes? May 20 17:57:28 pepere_: yes look into recipes/images May 20 17:57:39 ha thx! May 20 18:02:04 hi all. May 20 18:02:28 which web browsers do we have in openembedded?? May 20 18:02:40 .... May 20 18:05:56 boa cherokee apache nginx May 20 18:06:12 ha web browser sorry :) May 20 18:08:05 so.... May 20 18:09:00 I just seen there is konqueror May 20 18:09:20 it depends of your environment ... May 20 18:10:37 webkit/arora also May 20 18:13:57 fennec/firefox May 20 18:14:05 and old konqueror May 20 18:14:11 links May 20 18:14:12 w3m May 20 18:15:07 Spyzer: Is there one you're looking for in particular? Most of the popular ones are there May 20 18:16:24 actually i have to load a web based software named as svg edit here http://code.google.com/p/svg-edit/ !! This one loads properly in firefox and google chrome but not in konqueror and other qt based browsers. So i was just hoping for an embedded we browser which can run it. May 20 18:17:07 is fennec a fork of firefox?? May 20 18:18:45 is there a difference between base-image and minimal-image May 20 18:19:05 I've done the two, but the tar.bz2 does the same size... May 20 18:19:18 does someone knows the difference between the two? May 20 18:19:59 fennec is mobile-firefox May 20 18:25:29 Spyzer: try fennec, microb and minimo May 20 18:30:21 Jay7: it seems that firefox supports loading of an SVG as an object code but the others don't. Do u know any(embedded) browser which can handle this May 20 18:32:07 Spyzer: not sure May 20 18:32:44 i just read about microb and minimo . can these run on any embedded distro?? May 20 18:32:58 OE have recipes even May 20 18:33:09 i mean minimo's home page shows only windows CE and microb shows nokia based os May 20 18:33:11 but I'm unsure about svg support May 20 18:33:13 ohh ok May 20 18:33:20 hmmmm May 20 18:34:14 inkscape have support for svg ;) May 20 18:34:36 but it's not a browser :) May 20 18:34:41 he he, but i don't need that ;) May 20 18:35:03 * Jay7 afk for some time May 20 18:35:24 inkscape is cool May 20 18:36:39 Tartarus: on qemu ppc -fexpensive-optimizations or -frename-registers in FULL_OPTIMIZATION has bugs with gcc 4.4.4 May 20 18:40:32 k May 20 18:44:48 heh if I compare what Opt passes are enabled at Os and O2 strange enough they are same May 20 18:45:01 with one exception that -finline-functions in enabled at -Os May 20 18:45:05 thats a joke May 20 18:45:16 03Henning Heinold  07org.openembedded.dev * rf8175bcfbb 10openembedded.git/recipes/jamvm/jamvm_1.5.3.bb: jamvm: update to versiom 1.5.3 May 20 18:45:20 this is with gcc 4.4.4 May 20 18:45:53 Someone should write a testsuite thing that verifies that -Os binaries are actually smaller. May 20 18:46:57 broonie: I have done it in past May 20 18:47:08 it does reduce code size sometimes by 15% May 20 18:47:16 sometimes by 1% May 20 18:47:17 Yeah, it certainly used to for me. May 20 18:49:29 broonie: the manpages say it disables more optimizers at Os May 20 18:49:35 which makes sence May 20 18:52:19 hi jama May 20 18:55:55 ugh, 13 spots in the metadata where OVERRIDES is manipulated May 20 18:56:27 hm May 20 19:15:20 re florian May 20 19:15:31 can i have pyQt embedded install by bitbake??? May 20 19:15:47 spyzer? May 20 19:15:58 re May 20 19:15:58 pyqt with qt-embedded? May 20 19:16:39 i mean like i ahev pyQt for x11 can i have a s*similar* thing fro embeded systems as well through openembeded May 20 19:23:42 whe i install pyqt does the qte environment gets installed as well?? May 20 19:24:44 no May 20 19:24:53 It will install qt-X11 I think May 20 19:26:06 I've just created my own bibake recipes in /images/ but bitbake does not see it. Do I miss something, like refreshing a cache? May 20 19:26:07 woglinde_: can i have qte and corresponding pyqt for it?? May 20 19:26:46 pyqt is different from python-qt afaik, can i have qte with python-qt?? May 20 19:26:56 pepere_: its not parsed in may be May 20 19:27:05 woglinde_: hello May 20 19:27:22 hum ok, is parsing automatic? May 20 19:28:16 does openembedded hosts qt 4.7?? May 20 19:28:20 bitbake May 20 19:28:34 Spyzer: look into recipes May 20 19:29:00 Spyzer no qt 4.7 yet May 20 19:29:07 its beta by the way May 20 19:30:07 woglinde_: may u please tell me how can i use python with qte in any manner. PLEASE May 20 19:30:42 ok this si what i did May 20 19:30:49 so perhaps i malformed my recipes May 20 19:30:50 :) May 20 19:31:53 woglinde: ?? May 20 19:32:40 spyzer first oe recipes is serval versions behind upstream May 20 19:32:48 ups the pyqt-recipe May 20 19:33:25 second May 20 19:33:34 the recipe inherit qt4x11 May 20 19:33:43 so easiest way would be May 20 19:33:46 copy the recipe May 20 19:34:05 to python-pyqt-emb or something May 20 19:34:26 adn inherit qt4e May 20 19:34:42 and then try to compile May 20 19:34:52 ehhh bitbake May 20 19:34:57 the new recipe May 20 19:35:05 wow woglinde_ thanks a lot May 20 19:35:06 :) May 20 19:36:25 woglinde_: a bit more help. how can i tweak qte so as to point to svn of qte 4.7?? May 20 19:36:40 SRC_URI?? May 20 19:36:50 i suppose in qte-common.inc May 20 19:37:52 no May 20 19:38:22 then?? May 20 19:38:25 copy qt4-embedded_4.6.2.bb May 20 19:38:36 btw. qt is developed with git May 20 19:38:37 not svn May 20 19:40:34 copy qt4-embedded and then?? May 20 19:41:16 spyzer change the name to version you want May 20 19:41:26 port all necassry patches May 20 19:41:32 bitbake the new recipe May 20 19:46:08 woglinde_: lis python-qt4 different from pyqt and if it is can i install that from openembedded?? May 20 19:46:27 *sigh* May 20 19:46:42 find out some stuff by your self now May 20 19:46:49 sorry :( :( May 20 19:47:11 Anyone familar with the old psplash code about? May 20 19:47:19 and it seems the tasks invovled to getting pyqt with qt 4.7 is behind your actual knowledge May 20 19:47:22 lkml.org is back May 20 19:47:36 * Tartarus tries to reproduce the pstaging failure May 20 19:51:47 * khem has hacked up a script which removes everything thats needed to rebuild just the target and cross packages May 20 19:51:56 now I can build many machines May 20 19:52:50 PSTAGE_WORK is the variable you have to modify now, right/ May 20 19:53:10 bah May 20 19:53:32 * Tartarus sees no May 20 19:56:02 Tartarus: I need a second ack to bump up automake version can you add one :) May 20 20:01:02 bounce me the OE-devel msg pls, khem May 20 20:15:50 DJWillis, hi May 20 20:20:13 03Koen Kooi  07org.openembedded.dev * re0103e89b8 10openembedded.git/conf/distro/include/angstrom.inc: angstrom: fix java builds May 20 20:24:51 Tartarus: http://article.gmane.org/gmane.comp.handhelds.openembedded/32781 May 20 20:38:38 k, i'll just pw-am it for the 2nd one May 20 20:39:46 khem: OK with the email body being the commit msg? Or wanted something else May 20 20:42:41 yes May 20 20:42:46 nothing more nothing less May 20 20:44:00 03Khem Raj  07org.openembedded.dev * rdf6007e5d3 10openembedded.git/recipes/automake/automake_1.11.1.bb: (log message trimmed) May 20 20:44:00 automake_1.11.1.bb: Remove DEFAULT_PREFERENCE = "-1" May 20 20:44:00 Hi, May 20 20:44:00 I was working on upgrading gettext and some other packages to latest May 20 20:44:00 versions but they use features from automake 1.11. We have this recipe May 20 20:44:01 but it not enabled. I ask for approval to make it default. I have May 20 20:44:02 build few images thus far without any problems and SHR already uses May 20 20:55:19 hi May 20 20:57:10 http://tinderbox.openembedded.org/packages/575628/ coreutils-native Too many levels of symbolic links May 20 21:02:28 don't see where I got symbolic links in there. any pointers? May 20 21:07:42 JaMa: ping May 20 21:10:11 dcordes: is one of the directories in your build setup a link? May 20 21:11:05 florian, I don't think so. the oedir I'm using now is copied from a previous symlinkage May 20 21:11:20 but I cleared tmp May 20 21:11:58 dcordes: did u use cp -a May 20 21:12:05 dcordes: or cp --preserve ? May 20 21:12:59 cp -r /home/luke-g(this is the symlink dir)/blbalbla/oe/ /home/luke/build/ May 20 21:18:18 dcordes: that seems to be the problem May 20 21:18:26 it would copy the symlinks as such May 20 21:20:16 oops so my oedir is the symlink to old one May 20 21:20:55 /home/luke/build/openembedded/: directory May 20 21:20:56 dcordes: you should have done cp -rL May 20 21:21:04 no symlink May 20 21:21:37 dcordes: why dont you copy from the original location May 20 21:22:00 original beeing org.oe.dev? May 20 21:22:06 git May 20 21:22:41 no whereever the symlink was pointing to May 20 21:22:57 or you can do is find /home/luke/build/openembedded/ -type l -exec ls -l {} \; May 20 21:23:05 that will link any symbolic links May 20 21:23:10 What is the g++ option "-rpath-link"? May 20 21:23:11 btw. where is tmpdir May 20 21:23:21 khem, sorry think I missed something. May 20 21:23:44 IgorK: its a linker option man ld and read about it May 20 21:23:52 khem, I'm gonna backup conf and do a clean oedir May 20 21:24:14 and I was looking at "man gcc"... thanx khem May 20 21:25:12 got late for work just to see the result of the minimal image build. this is true love. May 20 21:27:54 is there an option to remove the path for an ld? May 20 21:29:48 s/image/distro/ May 20 21:38:51 Is there any option missing or needs to be added here: http://pastebin.com/sA0t1yEL May 20 21:39:05 ~seen Kensan May 20 21:39:10 for some reason linker couldn't find the library May 20 21:39:10 kensan is currently on #htc-linux #openmoko-cdevel. Has said a total of 113 messages. Is idling for 1d 58m 15s, last said: 'badcloud: if it's on the system the kernel module should be somewhere under /lib/modules/2.6.??'. May 20 21:50:01 aha aha May 20 22:00:04 hi, anyone around here who is familiar with crosscompiling qt? May 20 22:00:09 khem, ok I'm ready to go May 20 22:00:24 i need jpeg and svg support May 20 22:00:26 khem, just pulled and seen some chagnes in sane srcrevs but I think it doesn't affect me. anything to worry about ? May 20 22:20:03 Tartarus: I found this and I'm confused now...http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=394d321684eec271bc55438e3c01f74825d80988 May 20 22:21:37 Tartarus: oh, I read now your mail May 20 22:22:49 Hi. Trying to build boost 1.41 for omap3530evm. It first wants to build libxt, which fails looking for X11 headers. I see staging/i686-linux/usr/include/X11/Xfuncs.h, but I'm guessing I need headers in staging/arm. Any quick way to determine which X11 package I need to build first? May 20 22:25:30 dc May 20 22:29:43 hoj seems there is some deps missing May 20 22:30:02 hoj try bitbake libxt May 20 22:30:10 and than bitbake boost May 20 22:30:24 hm mom May 20 22:30:33 seems libxt seems to need some deps May 20 22:30:40 woglinde_: tried to build libxt and I get the same failure May 20 22:30:55 which header its need? May 20 22:33:15 Several: X11/Xfuncs.h X11/XlibConf.h X11/Xos.h. Also tried building libx11, which finished, but was not the right pkg. I thought maybe opkg on the x86 headers might tell me the native package, then I could build the cross version May 20 22:33:54 try bitbake xproto May 20 22:34:45 nope May 20 22:35:43 I probably cleaned some X11 stuff, but I would hope dependencies would rebuild what's needed May 20 22:47:54 03Andrea Adami  07org.openembedded.dev * r75bc8e0ea1 10openembedded.git/recipes/ (kexecboot/linux-kexecboot.inc linux/linux.inc): (log message trimmed) May 20 22:47:54 linux.inc: Rename getline in unifdef.c to parseline to avoid conflict. May 20 22:47:54 * move patch from linux-kexecboot.inc May 20 22:47:54 * see https://patchwork.kernel.org/patch/11166/ May 20 22:47:54 * committed in 2.6.29 (commit d15bd1067b1fcb2b7250d22bc0c7c7fea0b759f7) May 20 22:47:55 * Acked-by: Tom Rini May 20 22:47:56 * Acked-by: Henning Heinold May 20 23:56:28 03Henry von Tresckow  07org.openembedded.dev * rfedcbafe3b 10openembedded.git/recipes/linux/ (6 files in 2 dirs): May 20 23:56:28 linux-rp: make 2.6.26 compile for poodle May 20 23:56:28 * first try: kernel compiles May 20 23:56:28 * still hangs at: "Uncompressing Linux..... done, booting the kernel." May 20 23:56:28 Signed-off-by: Henry von Tresckow May 20 23:56:29 Acked-by: Andrea Adami May 21 01:06:33 What do I need to build in order to get a cross ldconfig? is it in binutils? May 21 01:16:18 "cross ldconfig"? May 21 01:16:32 why would you want that? May 21 01:17:19 ldconfig is part of glibc. does a cross ldconfig even exist? May 21 01:20:52 wouldn't really see the point. you can always run it on first boot on the device, but i doubt it's necessary in most cases. its likely the rpath will cover the default path May 21 01:32:10 yes there is cross ldconfig May 21 01:35:44 it looks like I need to update the ld search path on the host system... May 21 01:36:28 cat /home/arm/OE//home/arm/OE/temp/staging/armv4t-angstrom-linux-gnueabi/usr/lib.... May 21 01:37:21 which means that I can't pick up library from any other place May 21 01:37:43 AFAIK I need to run ldconfig May 21 01:38:17 or I just can put the path in the file mentioned? May 21 01:42:27 khem: "bitbake cross-ldconfig"? May 21 01:47:53 anybody? May 21 01:56:33 IgorK, i dont understand what problem you are trying to solve May 21 01:56:56 ldconfig updates the cache for the dynamic linker to find libraries May 21 01:57:22 grg: Yes, that's exactly my problem May 21 01:57:30 the dynamic linker isn't being run on your host, so how exactly is that a problem? May 21 01:58:06 I need the ld to search for "home/arm/OE/temp/staging/armv4t-angstrom-linux-gnueabi/usr/lib" and this path is not in the cache May 21 01:58:36 you are confusing compile time linking and runtime linking May 21 01:58:59 ldconfig's cache is for dynamic linking at runtime May 21 01:59:17 Im supplying the "-L" option to ld, but it comes back with :file not found May 21 01:59:24 ld links objects/executables to be run at a later date May 21 02:00:21 so why ld can't find the file? May 21 02:00:41 dunno May 21 02:00:51 i don't know what you're trying to do May 21 02:01:20 are you manually building something which needs to link against a lib in your OE staging area? May 21 02:02:17 grg: here it is: http://pastebin.com/QKwLG28x May 21 02:02:22 yes I am May 21 02:03:07 file is present but ld does not find it May 21 02:06:50 IgorK, you could run that command with strace, and see out where its trying to find libwx_gtk2_aui-2.8 May 21 02:07:15 i don't see anything obviously wrong with what you're trying to do May 21 02:08:52 IgorK, also, have you considered using pkg-config instead? May 21 02:12:09 e.g. PKG_CONFIG_SYSROOT_DIR=/home/arm/OE/temp/staging/armv4t-angstrom-linux-gnueabi --cflags --libs wxsomethingerutha May 21 02:28:19 so prepend that with an strace? do I need an cross strace? May 21 02:34:54 grg: also, it should go to "-L' path, right? May 21 02:52:52 grg: ping **** ENDING LOGGING AT Fri May 21 02:59:57 2010