**** BEGIN LOGGING AT Mon Aug 01 02:59:57 2011 Aug 01 08:26:39 khem: hey, stupid question. should __sync_fetch_and_add_4 be in libgcc_s? Aug 01 09:04:13 morning all Aug 01 09:12:59 hi bluelightning Aug 01 09:15:57 hi pb_ Aug 01 09:16:03 hello pb_ Aug 01 09:16:46 hi ant_work Aug 01 09:19:59 pb_: I see some oe-cre recipes added +INHIBIT_DEFAULT_DEPS = "1 . Is this a policy? Aug 01 09:20:19 I'm not quite sure what you mean. Aug 01 09:20:33 iirc, INHIBIT_DEFAULT_DEPS existed in oe.dev, and has done more or less forever Aug 01 09:20:34 some recipes doesn't need tooclhain (nothing to compile) Aug 01 09:20:54 oh right, you mean when the toolchain just isn't required Aug 01 09:21:28 seems reasonable enough to me, yeah. I don't know if it's actually an official policy or not Aug 01 09:21:33 I have tracked a serie of commits in oe-core adding that tovarious recipes Aug 01 09:21:48 is inmy to-do queue :) Aug 01 09:22:06 (as buying a new kb with working ) Aug 01 09:25:45 good morning Aug 01 13:22:16 pb_: heh, seems RP today is in a 'rejecting' mood ;) Aug 01 13:34:42 yeah, it seems Aug 01 13:34:42 very good Aug 01 14:12:13 03Koen Kooi  07org.openembedded.dev * r0b521d7bcf 10openembedded.git/classes/gtk-icon-cache.bbclass: Aug 01 14:12:14 gtk-icon-cache bbclass: only add runtime dependencies on hicolor-icon-theme when installing icons Aug 01 14:12:14 Also import a fix from OE-core for setting the loader dir. Aug 01 14:12:14 Tested with gnome-icon-theme and libsoup recipes on angstrom. Aug 01 14:12:14 Signed-off-by: Koen Kooi Aug 01 14:12:14 Acked-by: Steffen Sledz Aug 01 14:15:01 pb_: ehh, no, he's disappointed Aug 01 14:30:34 03Philip Balister  07master * rd7792dbfdc 10openembedded.git/recipes/uhd/ (3 files): Aug 01 14:30:34 uhd-firmware : Update recipe to package 3.2.1. Aug 01 14:30:34 Signed-off-by: Philip Balister Aug 01 14:30:40 03Philip Balister  07master * re8fe215d2c 10openembedded.git/recipes/uhd/ (uhd.inc uhd_git.bb): Aug 01 14:30:40 uhd : Update recipe to build 3.2.1 from git repo. Aug 01 14:30:40 Signed-off-by: Philip Balister Aug 01 14:30:40 03Philip Balister  07master * r3b1fa9d1f3 10openembedded.git/: Merge branch 'uhd-updates' Aug 01 14:34:59 03Philip Balister  07master * rb02759814b 10openembedded.git/recipes/asterisk/ (asterisk-1.4.39.2/init asterisk_1.4.39.2.bb): Aug 01 14:34:59 asterisk : Create log file directory in init script. Aug 01 14:34:59 If the log file is created on a tmpfs, the init script must the logfile Aug 01 14:34:59 directory and chown it to the asterisk user each time asterisk starts. Aug 01 14:34:59 Signed-off-by: Philip Balister Aug 01 14:37:35 uh uh Aug 01 14:37:47 rp is frustrated too Aug 01 14:53:44 woglinde: hi Aug 01 14:59:02 hi zecke Aug 01 15:00:07 hi zecke Aug 01 15:03:30 Morgan Aug 01 15:04:55 hi ka6sox Aug 01 15:05:19 hi ka6sox Aug 01 15:05:30 hi there, since we moved git to garnet things seem a lot more stable. Aug 01 15:05:35 morning woglinde Aug 01 15:37:00 hi joelagnel Aug 01 15:42:33 woglinde, gm Aug 01 15:43:33 opkg does not like the new architectures :( Aug 01 15:43:37 Package fbset version 2.1-r4 has no valid architecture, ignoring. Aug 01 15:43:53 Architecture: armv7a-vfp-neon Aug 01 15:44:35 fix it Aug 01 15:45:18 [OE-core] [PATCH 20/22] arch-armv7a.inc: fix armv7a-vfp-neon -> armv7a compat case ? Aug 01 15:45:39 hopefully all will stabilize soon Aug 01 15:47:06 03Joshua Lock  07master * rc9dd259243 10bitbake.git/lib/bb/ui/ (crumbs/configurator.py crumbs/hobprefs.py hob.py): Aug 01 15:47:07 hob: fix save/restore of toolchain preferences Aug 01 15:47:07 Add some (namespaced) custom variables to the configuration file for sake Aug 01 15:47:07 of this UI. Aug 01 15:47:07 Signed-off-by: Joshua Lock Aug 01 15:47:16 03Joshua Lock  07master * r68b992922b 10bitbake.git/lib/bb/ui/ (4 files in 2 dirs): (log message trimmed) Aug 01 15:47:16 hob: more reliable disabling of GPLv3 packages Aug 01 15:47:16 1. reflect GPLv3's presence in INCOMPATIBLE_LICENSE value in the UI Aug 01 15:47:16 The hob UI currently only supports GPLv3 as a value for Aug 01 15:47:17 INCOMPATIBLE_LICENSE but doesn't properly reflect whether the value is Aug 01 15:47:17 already set. This patch rectifies this. Aug 01 15:47:18 2. don't stomp over other INCOMPATIBLE_LICENSE values when disabling GPLv3 Aug 01 15:53:38 ensc|w: update, clean, try again Aug 01 16:11:06 ensc|w: that's weird. is that happening during rootfs construction or is this an on-target install? Aug 01 16:21:12 pb_: on target; e.g. 'opkg list-installed' Aug 01 16:21:58 ensc|w: sounds as though your /etc/opkg/arch.conf is somehow outdated Aug 01 16:23:26 packages for 'all' or 'phytec_pca102' architecture are listed Aug 01 16:24:08 do you touch PACKAGE_EXTRA_ARCHS in your conf files? Aug 01 16:24:44 till later Aug 01 16:25:11 I added armv7a-vfp-neon manually; perhaps I have to rebuild the package which provides etc/opkg/arch.conf Aug 01 16:25:24 unfortunately 'opkg search' does not work ;) Aug 01 16:28:10 rebuilding opkg-config-base fixed it; thx Aug 01 16:37:23 bbl Aug 01 18:03:39 guys i have a really strange error here Aug 01 18:03:43 http://pastebin.com/hS9Z4kxd Aug 01 18:04:30 i'm trying to bake linux-libc-headers and the qa_staging task fails with an IOError which shouldn't really matter at all Aug 01 18:04:42 what does libpng have to do with linux-libc-headers? Aug 01 18:06:26 Did you run out of disk space? Aug 01 18:09:17 half a terabyte free should be sufficient Aug 01 18:19:02 That is strange Aug 01 18:19:08 Not sure how in sync the dreambox bits are tho Aug 01 18:33:51 tartarus i've heard of others with the same exact repos where it built perfectly Aug 01 18:34:00 so something must have gone wrong here Aug 01 18:34:20 but unfortunately, cleaning linux-libc-headers and libpng doesn't help either :( Aug 01 18:52:58 i guess i have to cheat then -_- Aug 01 18:53:00 touch /dream/opendreambox-layers/tmp/stamps/mipsel-oe-linux/linux-libc-headers-2.6.37-r0.do_qa_staging Aug 01 18:54:07 -___- then it happens with some other packet Aug 01 18:54:51 http://pastebin.com/DsGx2ngp Aug 01 18:54:58 woglinde... do you have an idea what that could be? Aug 01 18:55:38 looks like libpng.la wasnt installed Aug 01 18:56:11 but I wonder why initscriptes needs libpng.ka Aug 01 18:56:13 ups .la Aug 01 18:56:39 yeah before i had the same issue with linux-libc-headers Aug 01 18:56:57 i touched the do_qa_staging Aug 01 18:57:26 dont know what you are touching Aug 01 18:57:31 try to build libpng again Aug 01 18:58:04 i cleaned libpng even Aug 01 19:02:27 you cleaned and didnt rebuild? Aug 01 19:02:30 so no wonder Aug 01 19:03:02 i thought that wouldn't matter unless linux-libc-headers was dependent on libpng? Aug 01 19:03:21 ... anyways ... this bitbake libpng is buidling a plethora of packages now Aug 01 19:03:38 i'm off to supper until then :) Aug 01 19:03:41 cu Aug 01 19:03:43 no sane class checks .la files Aug 01 19:03:46 all Aug 01 19:03:52 so you get the error Aug 01 19:09:16 if any smart people familiar with the gcc recipes could look at the following that would be great: http://lists.linuxtogo.org/pipermail/openembedded-core/2011-August/007342.html Aug 01 19:30:57 woglinde still working here Aug 01 20:10:09 I have a ping/pong scheme where a software-upgrade unpacks my OE-built rootfs.tar.bz2 from oe to an ext3 partition. Could I instead put a cpio archive in the partition, and run from an initramfs forever and ever? Aug 01 20:10:53 Well, initramfs means ram Aug 01 20:11:22 If you played some games to unpack onto the 'real' rootfs you might be able to survive using a lot of memory to old the rootfs to start with Aug 01 20:11:28 or maybe you've got a lotta memory :) Aug 01 20:11:36 or small fs Aug 01 20:13:49 I hope/think I can afford it. I have 256MB and I believe my tar.bz2 is ~57MB. Aug 01 20:20:37 well, off to sleep on it :) Aug 01 20:21:19 galak: thats advanced hackery to beat oe to use same sourcedir for building gcc in different flavours Aug 01 20:21:46 galak: interpreter paths are hardcoded into gcc itself Aug 01 20:21:47 khem: I think I've hit a bug wrt layers flattening & merging Aug 01 20:22:20 galak: so we need to have different interpreter paths for cross sdk Aug 01 20:22:30 it looks like if you add anew recipe, with higher PV, in a layer with lower priority, well, the new recipe is masked Aug 01 20:22:55 galak: and earlier when we had every gcc recipe have its own srctree then we could keep such hackery limited to those recipes e.g. gcc-cross-sdk etc. Aug 01 20:23:16 but if we have to merge and derive then from same source tree then there was this problem Aug 01 20:23:48 so at configure time we would emit a .h file which contains the prefix for ld.so Aug 01 20:24:11 and then we would concetenate that to name of ld.so to form the final path Aug 01 20:24:18 at gcc build time using defines Aug 01 20:24:46 this is GLIBC_DYNAMIC_LINKER Aug 01 20:25:18 anyhow so if you explain what issue you are seeing I would try to help you Aug 01 20:25:34 * khem was skeptical of using shared sources for gcc to start with Aug 01 20:25:49 there is this complexity vs. speed case always Aug 01 20:26:58 sometimes we should keep things simple too :) Aug 01 22:39:52 Can I have bitbake build an intermediate package for me so that I can see error messages with less stuff around them? Aug 01 22:40:14 For example, can I "bitbake glib-2.0" to see why "bitbake omap3-console-image" is failing? Aug 01 22:41:28 Also, I'm used to gentoo emerge. Is there anything like USE flags in bitbake? Aug 01 22:42:35 you can bitbake any recipe. Aug 01 22:42:37 images aren't special Aug 01 22:42:40 they're just recipes Aug 01 22:42:54 Hmm. Odd, didn't work before, but it worked just now. Aug 01 22:43:01 bitbake glib-2.0 finished without issue. Aug 01 22:43:16 bitbake gstreamer fails on do_configure, and says it can't find GLIB Aug 01 22:43:26 "| configure: No package 'glib-2.0' found" Aug 01 22:48:18 It seems it couldn't find glib-2.0.pc in the PKGCONFIG data folder, or something like that. Aug 01 22:49:32 Where is the sysroot in my project folder? Aug 01 22:49:55 build, tmp, or something else? Aug 01 22:50:49 alex@msp-vln-app03 ~/overo-oe/tmp/sysroots/armv7a-angstrom-linux-gnueabi $ find -iname glib-2.0.pc Aug 01 22:50:50 ./usr/lib/pkgconfig/glib-2.0.pc Aug 01 22:53:39 kergoth: How do I go about debugging why the configure is failing here? Aug 01 22:54:13 config.log might contain some more details Aug 01 22:54:30 config.log indicates that it couldn't find glib-2.0.pc Aug 01 22:54:59 it's trying to run the pkg-config from the sysroot Aug 01 22:55:21 /my_sysroot/usr/lib/pkgconfig/glib-2.0.pc is there. Aug 01 22:55:29 03Paul Menzel  07master * r033af22d56 10openembedded.git/recipes/xorg-app/iceauth_1.0.4.bb: (log message trimmed) Aug 01 22:55:30 iceauth: Improve `DESCRIPTION` and remove white space Aug 01 22:55:30 1. Improve description: s/a tool/Tool/. Therefore bump `PR`. Aug 01 22:55:30 2. Remove unneeded space when using `+=`. Aug 01 22:55:30 3. Please note that the fields `LICENSE` and `HOMEPAGE` are defined in the include file [2]. Aug 01 22:55:30 Based on a patch for meta-oe from Andreas [1]. Aug 01 22:55:31 [1] http://lists.linuxtogo.org/pipermail/openembedded-devel/2011-July/034168.html Aug 01 22:57:33 Oh well. I need to go home and unpack - just moved this weekend. **** ENDING LOGGING AT Tue Aug 02 02:59:57 2011