**** BEGIN LOGGING AT Tue Aug 09 02:59:56 2011 Aug 09 05:56:16 hello Aug 09 05:56:40 can anyone help me how to compile Mplayer with xvid support in oe.... Aug 09 07:49:44 Aug 09 09:25:13 Hi All, Please let me know SCST recipe is available in OE ? if not how can I get. Aug 09 09:45:19 Guest98586: look in the recipes dir if there is one, otherwise if you need it you must roll your own (see info on wiki and use existing recipes as examples) Aug 09 09:48:37 eFfeM_work: Thanks :) Aug 09 09:50:19 yw Aug 09 10:41:48 khem: hi, I have ICE in lttng-ust now with gcc-4.6.1, seems caused by linaro patches upto 07.2011 release as I haven't seen it before (happens only with armv7a here) Aug 09 10:43:17 khem: http://paste.pocoo.org/show/455210/ Aug 09 10:44:54 khem: vfprintf.s http://paste.pocoo.org/show/455211/ Aug 09 11:00:22 hi Aug 09 11:19:24 hello everybody, i habe a problem with the bitbake command :( Aug 09 11:20:01 when i try to compile "nano" i become following error: Aug 09 11:20:12 NameError: global name 'bb' is not defined Aug 09 12:11:14 khem: and ICE happens with -O1 -O2 -Os, builds fine only with -O0 Aug 09 12:23:01 JaMa: hi Aug 09 12:23:18 I was looking at meta-oe glib-2.0 recipe Aug 09 12:23:34 found come inconsistancy with openembedded recipe Aug 09 12:23:55 meta-oe recipe is DEPENDENT on dbus which is bringing x11 stuff Aug 09 12:24:20 but openembedded glib-2.0 recipe is not debendent on dbus Aug 09 12:26:25 DEPENDS = "glib-2.0-native gtk-doc-native zip dbus" .... line from glib.inc file which is being included in glib-2.0_git.bb recipe in meta oe Aug 09 12:26:49 shouldn't we drop dbus over here Aug 09 12:28:44 Noor: http://git.gnome.org/browse/glib/tree/configure.ac shows dbus for GDBus test cases, but it's optional.. Aug 09 12:39:44 JaMa, Noor, probably preferable would be to split x11 parts for runtime packages (if it isn't the case already). So that x11 parts will be built, but not installed for e.g. console-image Aug 09 12:44:15 I was just building udev from openembedded and meta-oe/oe-core and found that there is a lot of difference in the total tasks Aug 09 12:44:57 in meta-oe/oe-core setup the total task were around 1010 to build udev and in openembedded it was around 800 Aug 09 12:45:25 so when I looked it I found that dbus is the extra thing that is added in glibc-2.0 Aug 09 12:45:43 and it is bringing around 200 extra tasks Aug 09 12:46:14 so just wondering that we really need dbus in glib recipe Aug 09 12:46:43 lumag: GDBus tests are not part of libglib-2.0-0 and dbus is not in rdepends.. Aug 09 12:46:49 in openembded the glib configure complains about that I dont have dbus but it builds and work fine Aug 09 12:48:09 yeah its is not in rdepends.... just looking to reduce the build packages Aug 09 12:48:43 Noor: btw this was causing also dependency loop.. Aug 09 12:50:34 Jama: dependency loop? .... i was able to build udev after removing dbus from glib ... may be it showed up in other package building Aug 09 12:52:00 Noor: http://comments.gmane.org/gmane.comp.handhelds.openembedded/40368 Aug 09 12:58:23 JaMa: ok the loop between dbus and glib-2.0 Aug 09 14:22:06 03Joshua Lock  07master * r46953d3b45 10bitbake.git/lib/bb/ui/hob.py: (log message trimmed) Aug 09 14:22:06 bb/ui/hob: move some code around to avert a race Aug 09 14:22:06 The data-generated and model-updated signals are different, the model Aug 09 14:22:06 should only be accessed *after* the model-updated signal. Move code Aug 09 14:22:06 setting the image combo's backing model to the model-updated callback to Aug 09 14:22:07 ensure the combo is accurately set when changing the machine with an Aug 09 14:22:07 image selected. Aug 09 14:22:16 03Joshua Lock  07master * r0ed58f28e8 10bitbake.git/lib/bb/ui/crumbs/tasklistmodel.py: (log message trimmed) Aug 09 14:22:16 bb/ui/crumbs/tasklistmodel: prevent hang when removing item Aug 09 14:22:16 It's possible to trigger an infinite recursion when removing a package Aug 09 14:22:16 where many of the dependencies share their dependencies. Prevent this by Aug 09 14:22:17 keeping a list of removed item names and only removing the item when it's Aug 09 14:22:17 not in the list. Aug 09 14:22:18 Addresses [YOCTO #1319] Aug 09 14:22:18 03Jessica Zhang  07master * rc3827690b9 10bitbake.git/lib/bb/ui/hob.py: Aug 09 14:22:19 bb/ui/hob: Restore toolchain relevant preference settings for build Aug 09 14:22:19 Fixes [#YOCTO 1354] Aug 09 14:22:20 Signed-off-by: Jessica Zhang Aug 09 14:22:20 Signed-off-by: Joshua Lock Aug 09 14:32:15 03Philip Balister  07master * r10a36a519c 10openembedded.git/recipes/gnuradio/ (gnuradio.inc gnuradio_git.bb): Aug 09 14:32:15 gnuradio : Bump SRCREV to current version. Fix pacakging warning. Aug 09 14:32:15 Signed-off-by: Philip Balister Aug 09 14:46:51 hello, i'm following the instructions on http://gumstix.org/create-a-bootable-microsd-card.html but i can't find any file called uImage.bin. does anyone know where this would be? Aug 09 15:51:32 RP__:around ? Aug 09 15:51:48 galak: yes Aug 09 15:52:16 RP__: so the telepathy-python issue, I think its a more generic python and libdir issue Aug 09 15:52:49 galak: quite likely. We've got so far addressing those issues but not got everything fixed :( Aug 09 15:57:22 but I dont think what's be done for other python packages is working Aug 09 15:57:39 for example: Aug 09 15:57:40 commit 16e29ba77ddcf0320cd9ce552650bd69761d0bf0 Aug 09 15:57:40 Author: Dongxiao Xu Aug 09 15:57:40 Date: Wed Jul 13 16:43:44 2011 +0800 Aug 09 15:57:41 xcb-proto: add multilib support. Aug 09 15:58:33 still ends up with libs in /usr/lib: Aug 09 15:58:33 WARNING: /usr/lib/python2.6/site-packages/xcbgen/__init__.pyo Aug 09 15:58:42 WARNING: /usr/lib/python2.6/site-packages/xcbgen/matcher.py Aug 09 15:58:42 WARNING: /usr/lib/python2.6/site-packages/xcbgen/matcher.py Aug 09 15:58:46 etc.. Aug 09 16:21:04 galak: we need to work out why then... Aug 09 16:21:45 looks like we regen the aclocal.m4 and re-uses python template from automake Aug 09 16:21:46 that has: Aug 09 16:22:11 'lib' hardcoded: Aug 09 16:22:11 am_cv_python_pythondir=$PYTHON_PREFIX/lib/python$PYTHON_VERSION/site-packages Aug 09 16:23:03 so not sure if best solution is to patch automake's m4/python.m4 or what at this point Aug 09 16:29:01 is TARGET_ARCH still used? it's explicitly unset in my kernel build… perhaps replaced by just ARCH? TRANSLATED_TARGET_ARCH? TUNE_ARCH? Aug 09 16:31:22 RP__: so we could do something like the following in meta/site/common-glibc: Aug 09 16:31:26 # python Aug 09 16:31:26 am_cv_python_pyexecdir=${libdir}/python2.6/site-packages Aug 09 16:31:26 am_cv_python_pythondir=${libdir}/python2.6/site-packages Aug 09 16:36:25 however, this seems wrong because of the hardcoded python version Aug 09 16:37:03 does anyone know how to rename ethernet interfaces in the kernel? Aug 09 16:37:38 toobluesc_: you don't rename them in kernel, you can use udev rules to do it Aug 09 16:38:12 galak, I need to rename the ethernet devices before I mount my rootfs, hence udev isn't started yet Aug 09 16:38:49 hmm, not sure what to suggest. Why doesnt ethX work? Aug 09 16:40:05 my original interface is eth0 Aug 09 16:40:19 which works fine with u-boot kernel etc Aug 09 16:40:25 i've added a driver and another chip Aug 09 16:40:42 but it ends up booting and moving eth0 -> eth1_rename Aug 09 16:40:50 and the other chip somehow becomes eth0 Aug 09 16:41:03 just want to force the order of enumeration somehow Aug 09 16:41:41 I thought that the order I added the platform devices would be sufficient, but no long yet Aug 09 16:46:17 drivers register in the desired order and the platform devices are added in the correct order Aug 09 16:46:23 just somehow it always stomps on the first one Aug 09 16:59:43 galak: we could export those in one of the python classes I guess Aug 09 17:05:55 RP__: can you send a pointer to what classes to look at and I'll see what can be done Aug 09 17:06:13 RP__: respond to email if you dont mind as I'll be dropping off IRC in a sec Aug 09 17:54:01 galak, i figured it out Aug 09 17:54:11 toobluesc_: how? Aug 09 17:54:36 reason the ordering was screwy is because the davinci_emac init code was a late_initcall() rather the module_init() Aug 09 17:54:44 and the smsc911x driver was using module_init() Aug 09 17:54:57 so smsc911x always won Aug 09 17:54:59 and got eth0 Aug 09 17:55:17 and then the davinic_mac code isn't written correctly to use the ethernet auto enumeration code :-/ Aug 09 17:55:36 but all is well now Aug 09 18:08:49 RP__: still around? Aug 09 18:21:12 hmm, is there a known problem with OE where things like coreutils, m4, autoconf etc all end up linking GNUmakefile to itself instead of Makefile? Aug 09 19:30:57 galak: kind of Aug 09 19:33:00 RP__: posted patches Aug 09 19:34:07 is there a default username/password for the omap3 console image? Aug 09 20:29:54 RA_drc: root with no password Aug 09 20:30:03 housel: thanks Aug 09 20:57:58 JaMa: ok. Can you send the preprocessed file my way ? Aug 09 21:11:39 khem: preprocessed by -S? Aug 09 21:15:53 khem: vfprintf.E http://paste.pocoo.org/show/455583/ Aug 09 21:19:09 thanks JaMa|Zzz Aug 09 22:23:15 hello, do you guys know if OE could work on Asus routers that currently run OpenWRT? Aug 10 00:11:23 hmm, is there a known problem with OE where things like coreutils, m4, autoconf etc all end up linking GNUmakefile to itself instead of Makefile? Aug 10 02:04:33 tzanger: what do u mean ? Aug 10 02:04:46 tzanger: are you trying it on BSD Aug 10 02:07:41 nope not at all Aug 10 02:07:52 just a regular linux system, ubuntu 10.10 Aug 10 02:10:21 bitbake starts up, grabs the file, expands it, runs configure, then explodes because of an infinite recursion (GNUmakefile links to GNUmakefile). If I delete it and link it to Makefile, things go well Aug 10 02:10:26 at least until autoconf Aug 10 02:57:51 tzanger: hmmm Aug 10 02:58:16 the do_configure log specifically says it's linking GNUmakefile to GNUmakefile Aug 10 02:58:28 so it's not zigging when it shoudl be zagging, it really thinks it should be doing that Aug 10 02:59:15 e.g. Aug 10 02:59:18 config.status: linking /home/andrew/angstrom/x86/angstrom/build/tmp-angstrom_2008_1/work/x86_64-linux/autoconf-native-2.65-r12.1/autoconf-2.65/GNUmakefile to GNUmakefile Aug 10 02:59:20 Do you have some symlinks in your build path? Aug 10 02:59:28 hm Aug 10 02:59:33 yes Aug 10 02:59:41 Bingo Aug 10 02:59:43 Don't Aug 10 02:59:44 /home/andrew/angstrom links to /mnt/foo Aug 10 02:59:50 really, that breaks this? **** ENDING LOGGING AT Wed Aug 10 02:59:57 2011