**** BEGIN LOGGING AT Fri Oct 22 02:59:58 2010 Oct 22 03:05:40 hmpf.. built an angstrom image for zaurus 5600, updater won't take it, 2MB too large for internal flash :( Oct 22 03:16:02 03Chris Larson  07master * r18690b2e5a 10openembedded.git/classes/patch.bbclass: Oct 22 03:16:02 patch.bbclass: avoid traceback for CmdError at resolve Oct 22 03:16:02 Signed-off-by: Chris Larson Oct 22 03:16:07 03Chris Larson  07master * re84f4995e2 10openembedded.git/recipes/mpfr/ (mpfr-3.0.0/p3.patch mpfr-3.0.0/p4.patch mpfr_3.0.0.bb): Oct 22 03:16:07 mpfr: update from p3 to p4 patchset Oct 22 03:16:07 Signed-off-by: Chris Larson Oct 22 03:16:07 03Chris Larson  07master * r7e4958f03e 10openembedded.git/lib/oe/patch.py: Oct 22 03:16:08 oe.patch: fix 'd' references Oct 22 03:16:08 Signed-off-by: Chris Larson Oct 22 03:16:14 03Chris Larson  07master * rc7a68fe6ff 10openembedded.git/lib/oe/patch.py: Oct 22 03:16:14 oe.patch: use oe.process Oct 22 03:16:14 Signed-off-by: Chris Larson Oct 22 03:16:14 03Chris Larson  07master * r44887999fa 10openembedded.git/lib/oe/patch.py: Oct 22 03:16:15 oe.patch: don't use dry-run for 'patch' Oct 22 03:16:15 There are patches which 'patch' thinks it can't apply, but which do apply, so Oct 22 03:16:16 stop using dry-run. In addition, --dry-run isn't compatible with all versions Oct 22 03:16:16 of patch. Oct 22 03:16:17 Signed-off-by: Chris Larson Oct 22 03:16:19 03Chris Larson  07master * rd1d6f80d4d 10openembedded.git/classes/utils.bbclass: Oct 22 03:16:19 oe_libinstall: handle .dylib Oct 22 03:16:19 Signed-off-by: Chris Larson Oct 22 03:16:19 03Jason Kridner  07master * rbe512828d0 10openembedded.git/recipes/quilt/quilt-native.inc: quilt-native: disable getopt for darwin buildhosts Oct 22 03:16:20 03Jason Kridner  07master * r047ef9b9e2 10openembedded.git/recipes/openssl/openssl.inc: Oct 22 03:16:21 openssl: add i386-darwin to known buildhosts Oct 22 03:16:21 Signed-off-by: Jason Kridner Oct 22 03:16:22 Signed-off-by: Chris Larson Oct 22 03:16:22 03Chris Larson  07master * r1c164de88d 10openembedded.git/lib/oe/process.py: Oct 22 03:36:42 03Chris Larson  07org.openembedded.dev * rcaa6067f7f 10openembedded.git/ (58 files in 41 dirs): Various fixes for cp argument portability Oct 22 03:36:42 03Chris Larson  07org.openembedded.dev * rb79a20523f 10openembedded.git/recipes/readline/ (darwin.patch readline_5.2.bb): Oct 22 03:36:42 readline-5.2: fix for darwin builds Oct 22 03:36:42 Signed-off-by: Chris Larson Oct 22 03:36:42 03Chris Larson  07org.openembedded.dev * r788c62604c 10openembedded.git/bin/ (cp sed): Oct 22 03:36:43 bin/{cp,sed}: Add shell script wrappers for portability Oct 22 03:36:44 Signed-off-by: Chris Larson Oct 22 03:36:44 03Chris Larson  07org.openembedded.dev * r39fae5b646 10openembedded.git/conf/build/ (3 files in 2 dirs): Oct 22 03:36:45 conf/build: darwin fixups - COMMAND_MODE, -m32, PATH Oct 22 03:36:50 - For ppc darwin, use the wrapper scripts Oct 22 03:36:50 - For i386 darwin, set -m32, as uname shows i386 even on 64 bit darwin Oct 22 03:36:50 - For i386 and ppc darwin, set COMMAND_MODE to fix incompatible shell utility Oct 22 03:36:50 behavior, for example 'echo' not handling -n as expected Oct 22 03:36:50 Signed-off-by: Chris Larson Oct 22 03:36:50 03Chris Larson  07org.openembedded.dev * r1f94eb989a 10openembedded.git/recipes/openssl/openssl.inc: Oct 22 03:36:51 openssl: kill version script for darwin Oct 22 03:36:51 This might not be the ideal solution, but it does get openssl-native to build Oct 22 03:36:52 on a darwin machine. Oct 22 03:36:53 Signed-off-by: Chris Larson Oct 22 03:36:53 03Koen Kooi  07org.openembedded.dev * r4060ff0edc 10openembedded.git/recipes/mpfr/mpfr.inc: mpfr: remove 'PATCHES' file so it can build on case insensitive filesystems where quilt wants to creat the 'patches' dir Oct 22 03:36:54 03Chris Larson  07org.openembedded.dev * rfa83629f8e 10openembedded.git/conf/build/i386-darwin.conf: Oct 22 06:20:25 gm Oct 22 06:22:29 eFfeM_work, morning Oct 22 07:05:54 JaMa|W: I just posted a patchset which should fix that binutils issue you reported. but you have to rebuild from scratch Oct 22 07:06:09 unfortunately :/ Oct 22 07:06:47 JaMa|W: try it out and let me know if you see any issues Oct 22 07:07:44 yup seen it.. will start rebuild soon Oct 22 07:07:51 cool thx Oct 22 07:09:54 khem: I'll rebuild from scratch, but can you explain a bit why it's needed now? I guess .la format is not changed right? Oct 22 07:10:18 asking just because on shr buildhost rebuild for all supported archs takes about 10 days (slow VM).. Oct 22 07:10:21 JaMa|W: well it changes the content of libtool scripts Oct 22 07:11:17 JaMa|W: 10 days !! Oct 22 07:11:27 man thats too much Oct 22 07:11:54 khem: well we build for armv[467] now and ie qt4 takes 6hours there (and 1 hour on my host) Oct 22 07:11:56 and you always use bleeding edge Oct 22 07:12:08 khem: but that's only donated server we have.. Oct 22 07:12:19 JaMa|W: ok Oct 22 07:13:12 JaMa|W: I wanted to have an image which boots in qemu and has qt Oct 22 07:13:18 so which image should I try Oct 22 07:13:42 we're building shr-images and qt is only in task-shr-feed Oct 22 07:13:58 shr-image is based of on qt Oct 22 07:14:20 can I build this image with minimal distro I wonder Oct 22 07:14:22 no.. shr-images have mostly all gui in efl Oct 22 07:14:55 some parts are using (python)gtk, but nothing in image needs qt Oct 22 07:15:18 JaMa|W: so why do u waste qt4 Oct 22 07:15:35 building time Oct 22 07:15:39 it's in feed for extra packages users are able to install Oct 22 07:16:19 ok Oct 22 07:16:23 i mean't task-shr-feed is not in images rdepends, but users are used to those packages available after rebuild Oct 22 07:16:57 I would like to use a qt4-image or something Oct 22 07:17:07 which boots into qt based GUI Oct 22 07:18:50 there was Qt based gui for freerunner, but not sure if it was built by some out-of-tree OE based http://wiki.openmoko.org/wiki/Qalee Oct 22 07:19:49 but it seems dead for a while :/ Oct 22 07:21:31 alright Oct 22 07:21:33 time to sleep now Oct 22 07:21:36 gn Oct 22 07:46:35 CrazyFoam: ping Oct 22 08:30:09 03Koen Kooi  07org.openembedded.dev * r6394bcb5c3 10openembedded.git/recipes/ti/ti-dmai_svn.bb: Oct 22 08:30:09 ti-dmai: bump SRCREV for omapl138 Oct 22 08:30:09 * this fixes the codecerver changes for c6accel Oct 22 08:30:09 Signed-off-by: Koen Kooi Oct 22 08:30:14 03Koen Kooi  07org.openembedded.dev * r591ca10af9 10openembedded.git/recipes/ti/gstreamer-ti.inc: Oct 22 08:30:14 gstreamer-ti: bump PR for dmai change Oct 22 08:30:14 Signed-off-by: Koen Kooi Oct 22 08:40:51 good morning Oct 22 08:41:33 good morning Oct 22 09:04:14 eFfeM_work: all 7 patches looks also good to me Oct 22 09:04:30 thanks Oct 22 09:05:17 they seem trivial, but it seems all these recipes have a maintainer so I expected him/her to ack it Oct 22 09:05:28 especially since I do not build or use things like mono Oct 22 09:06:44 eFfeM_work: you can store -e output and confirm that with that patch it's the same.. I've sent ack to ML Oct 22 09:07:20 JaMa|W: ah ok, didn't think of that (actually I hardly use -e) Oct 22 09:07:45 i'm more concerned about some of the other empty strings Oct 22 09:08:11 and those empty strings I guess are sometimes used to overwrite ie EXTRA_OECONF set by their .inc file (but haven't checked myself) Oct 22 09:09:23 yes, I used "" in that way in a couple of recipes Oct 22 09:10:15 yes, that definitely needs to be verified (but then I'd say the recipe still would require fixing, because then the .inc file should not set it, only real reason I know to do it is because in virtclass_native you want to nullify the target value Oct 22 09:10:29 * eFfeM_work still thinks we should have _target suffix Oct 22 09:12:54 eFfeM_work: surely in the case of sensible vars more investigation is needed Oct 22 09:13:05 treue Oct 22 09:13:08 true Oct 22 09:13:34 issue might also be that some of those vars (like LDFLAGS) might be set in a class Oct 22 09:13:39 iirc most of the time I had to unset some IMAGE vars Oct 22 09:13:58 yep, IMAGE_*_MANAGER e.g. Oct 22 09:14:13 that's why I was hoping for some reply from people that are more knowledgeable about what constructs are definitely sound Oct 22 09:14:24 because inheriting image brings lot of stuff Oct 22 09:14:59 I can't say for the specific cases you noticed, I'll have a look when possible Oct 22 09:15:22 ant_work: would be cool Oct 22 09:15:52 guess the EXTRA_* things are empty when you start. same for SRC_URI Oct 22 09:16:47 well, building two images together (initramfs) we did need EXTRA_IMAGEDEPENDS = "" Oct 22 09:17:00 in the second one Oct 22 09:17:20 but sure, at the beginning should be empty Oct 22 09:17:22 ah ok, never did that Oct 22 09:18:15 but as JaMa|W stated, a bitbake -e test which does not reveal differences will make clear it is a sound patch Oct 22 09:23:44 mainly wondering about the recipe that resets RDEPENDS but has a DEPENDS, the one with PROVIDES="" etc Oct 22 09:30:44 03Martin Jansa  07master * rafc9616987 10openembedded.git/recipes/freesmartphone/cornucopia.inc: Oct 22 09:30:45 cornucopia: bump SRCREV for fixed build of fsogsmd on pre Oct 22 09:30:45 Signed-off-by: Martin Jansa Oct 22 09:30:55 03Martin Jansa  07master * r2f417a61ef 10openembedded.git/recipes/freesmartphone/msmcomm.inc: Oct 22 09:30:55 msmcomm: bump SRCREV Oct 22 09:30:55 Signed-off-by: Martin Jansa Oct 22 10:10:30 morning Oct 22 10:10:39 hi mickey|office Oct 22 10:27:40 Good moring, florian! Oct 22 10:28:24 hi mfhk Oct 22 10:28:29 hi mickey|office Oct 22 10:28:57 I have done quite a bit of changes to your mini6410 configs. Now, should I setup my local overlay ... Oct 22 10:29:22 mfhk: what did you change? Oct 22 10:30:02 because i have OK6410 , I need to change LCD driver, and Oct 22 10:30:30 ubifs commands for 1GB NAND , Oct 22 10:30:37 and NAND drivers Oct 22 10:30:58 What is the best practice for this situation? Oct 22 10:32:37 sounds like you need to define it as a new machine as its significantly different Oct 22 10:33:32 Xora, should I define this new machine in the original OE tree? Oct 22 10:33:45 Or, setup local overlay? Oct 22 10:34:00 mfhk: test it in an overlay then submit a patch when it all works to OE Oct 22 10:34:21 although the current OEDEM RFC might change that Oct 22 10:34:43 I see. Oct 22 10:34:43 hi mickey|office, florian Oct 22 10:34:50 Thanks! :) Oct 22 10:34:50 hi pb_ Oct 22 10:39:08 eFfeM_work: did you use -b with -e? default SRCREV is assigned in bitbake.conf (unconditionaly on SRC_URI) Oct 22 10:39:45 03Martin Jansa  07master * r3f307f24f6 10openembedded.git/recipes/lvm2/ (7 files in 4 dirs): Oct 22 10:39:45 lvm2: add patch for build with newer eglibc, drop empty do_stage() Oct 22 10:39:45 * also rename files to lvm2 for faster FILESPATH lookup Oct 22 10:39:45 Signed-off-by: Martin Jansa Oct 22 11:12:34 JaMa|W: yes Oct 22 11:15:53 & did overlook theone in bitbake.conf Oct 22 11:16:12 yes = yes, I used -b Oct 22 11:22:39 hey pb_ , hi florian Oct 22 12:09:54 Hi, is there a package on OE to make general benchmark (cpu, io, memory bandwith) on a system ? Oct 22 13:12:12 03Michael Smith  07master * r1e47e9629e 10openembedded.git/recipes/gcc/ (gcc-4.3.3.inc gcc-4.3.3/gcc-4.3-backport-print-sysroot.patch): Oct 22 13:12:12 gcc 4.3.3: backport -print-sysroot for libtool Oct 22 13:12:12 libtoolized configure scripts need this in order to autodetect sysroot Oct 22 13:12:12 path when --with-sysroot is passed. Oct 22 13:12:12 Signed-off-by: Michael Smith Oct 22 13:12:13 Acked-by: Khem Raj Oct 22 13:12:26 03Michael Smith  07master * rfb2c1fb732 10openembedded.git/recipes/gcc/ (gcc-4.3.4.inc gcc-4.3.4/gcc-4.3-backport-print-sysroot.patch): Oct 22 13:12:26 gcc 4.3.4: backport -print-sysroot for libtool Oct 22 13:12:26 libtoolized configure scripts need this in order to autodetect sysroot Oct 22 13:12:26 path when --with-sysroot is passed. Oct 22 13:12:27 Signed-off-by: Michael Smith Oct 22 13:12:27 Acked-by: Khem Raj Oct 22 13:37:28 I've copied an image from gusmtix and have the user creation/ on screen keyboard login thing Oct 22 13:37:41 anyone know how to make this appropriate for a real monitor? Oct 22 13:46:39 03Michael Smith  07master * r5cbe32ec7e 10openembedded.git/recipes/jack/jack_0.118.0.bb: Oct 22 13:46:39 jack: temporarily use Angstrom source mirror Oct 22 13:46:39 Upstream tarball has gone missing. Oct 22 13:46:39 Signed-off-by: Michael Smith Oct 22 13:47:13 Tryum: try unixbench Oct 22 14:04:59 03Martin Jansa  07master * r637f98d8b5 10openembedded.git/recipes/shr/shr-specs_git.bb: Oct 22 14:04:59 shr-specs: remove old staging Oct 22 14:04:59 Signed-off-by: Martin Jansa Oct 22 14:36:29 morning Oct 22 14:41:01 hmm Oct 22 14:43:23 morning Oct 22 14:43:42 kergoth: are you working on building OE under mac os X, or building some kind of BSD under os X? Oct 22 14:43:59 building OE on my macbook, i386-darwin Oct 22 14:44:05 have yet to successfully complete a task-boot, though Oct 22 14:44:07 its slow going Oct 22 14:44:10 what machine? Oct 22 14:44:12 target i mean Oct 22 14:44:25 not particularly relevent, i pick whatever i expect to actually build Oct 22 14:44:30 at the moment i think its beagleboard Oct 22 14:44:35 slick Oct 22 14:44:56 whats a pain of course, is unrelated issues breaking the testing.. maybe i should be using one of the testing branches instead of master :) Oct 22 15:16:33 03Khem Raj  07master * re12debc8e8 10openembedded.git/recipes/binutils/ (binutils-2.20.1/libtool-2.4-update.patch binutils.inc): Oct 22 15:16:33 binutils: Update libtool macros to reflect --with-sysroot -> --with-libtool-sysroot rename Oct 22 15:16:33 * We dont need to filter out with-sysroot anymore Oct 22 15:16:33 Signed-off-by: Khem Raj Oct 22 15:16:33 Acked-by: Chris Larson Oct 22 15:16:33 Acked-by: Koen Kooi Oct 22 15:16:36 03Khem Raj  07master * r46d563f4bb 10openembedded.git/classes/autotools.bbclass: Oct 22 15:16:36 autotools.bbclass: Use --with-libtool-sysroot instead of --with-sysroot Oct 22 15:16:36 Signed-off-by: Khem Raj Oct 22 15:16:36 Acked-by: Chris Larson Oct 22 15:16:37 Acked-by: Koen Kooi Oct 22 15:16:51 03Khem Raj  07master * r7cfd4da583 10openembedded.git/recipes/openldap/openldap_2.4.23.bb: Oct 22 15:16:51 openldap_2.4.23.bb: Fetch ltmain.sh from the new location in staging Oct 22 15:16:51 Signed-off-by: Khem Raj Oct 22 15:16:51 Acked-by: Chris Larson Oct 22 15:16:51 Acked-by: Koen Kooi Oct 22 15:16:51 03Khem Raj  07master * re0eabdaf07 10openembedded.git/recipes/gcc/ (gcc-4.5.inc gcc-configure-target.inc): Oct 22 15:16:52 gcc-configure-target.inc: --with-sysroot -> --with-libtool-sysroot rename means we dont need to filter out --with-sysroot Oct 22 15:16:54 Signed-off-by: Khem Raj Oct 22 15:16:54 Acked-by: Chris Larson Oct 22 15:16:54 Acked-by: Koen Kooi Oct 22 15:17:01 03Khem Raj  07master * r6f69d22a7b 10openembedded.git/recipes/libtool/ (5 files in 2 dirs): (log message trimmed) Oct 22 15:17:01 libtool: Rename --with-sysroot to --with-libtool-sysroot Oct 22 15:17:01 * To avoid name conflict with binutils and gcc we rename the option Oct 22 15:17:01 * We also stage the cross and native libtool in same relative path Oct 22 15:17:01 so it can be accessed equally well from recipes using BBCLASSEXTEND = "native" Oct 22 15:17:01 Signed-off-by: Khem Raj Oct 22 15:17:01 Acked-by: Chris Larson Oct 22 15:30:05 oh hell. Oct 22 15:30:14 eglibc svn has case sensitivity issues Oct 22 15:30:15 bite me Oct 22 15:47:53 heyho Oct 22 15:57:15 ugh, gnutls-native fails to build entirely if you use gettext-native 0.17 Oct 22 15:57:21 we need to drop 0.17, or fix those Oct 22 15:57:28 * kergoth grabs for 0.17 pins Oct 22 15:57:32 s/grabs/greps/ Oct 22 16:12:48 * kergoth creates a case sensitive disk image to try another oe build in Oct 22 16:13:01 (temporary) Oct 22 16:13:28 * kergoth also rebases the darwin fix commits upon the testing branch Oct 22 16:13:36 god i love git-new-workdir Oct 22 16:51:26 damn, why is tar compatibility such a pain Oct 22 16:51:40 gnu can use --owner, but afaict you have to use an mtree file to adjust that with osx's Oct 22 16:51:48 and of course, gnu doesn't seem to support mtree files Oct 22 16:55:19 kergoth: ping Oct 22 16:55:24 morphis: pong Oct 22 16:55:54 kergoth: I don't know if playya already told you, but please can you pull for tslib from http://github.com/playya/tslib:master? Oct 22 16:56:14 i've been pulling from the for-upstream branch up till now Oct 22 16:56:19 i can pull from master if you'd like Oct 22 16:56:21 hold Oct 22 16:56:21 ah great Oct 22 16:56:29 you already pulled Oct 22 16:56:43 gm Oct 22 16:57:18 kergoth: which branch from playya/tslib do you merged? Oct 22 16:57:28 the for-upstream one? Oct 22 16:57:34 i merged for-upstream recently Oct 22 16:57:39 hm ok Oct 22 16:57:41 since i am upstream, i figured that was its purpose :) Oct 22 16:57:51 it was Oct 22 16:58:22 looks like an attempt to merge from your current master results in conflicts, so if you could handle that merge by updating to current upstream master, i'd be willing to pull from your master Oct 22 16:58:22 hm, than I have to rebase master on for-upstream so you can pull again Oct 22 16:58:25 * kergoth nods Oct 22 17:05:33 okay, screw this help2man dependency, i'm fixing the ones i can find Oct 22 17:07:14 03Frans Meulenbroeks  07org.openembedded.dev * r0e9ffb16a0 10openembedded.git/recipes/mono/mono-mcs-intermediate_2.6.3.bb: Oct 22 17:07:14 mono-mcs-intermediate: remove useless SRC_URI += "" Oct 22 17:07:14 Signed-off-by: Frans Meulenbroeks Oct 22 17:07:14 Acked-by: Martin Jansa Oct 22 17:07:14 Acked-by: Chris Larson Oct 22 17:07:17 03Frans Meulenbroeks  07org.openembedded.dev * r3f66827ded 10openembedded.git/recipes/avahi/avahi.inc: Oct 22 17:07:17 avahi.inc: remove RDEPENDS_append = "" Oct 22 17:07:18 Signed-off-by: Frans Meulenbroeks Oct 22 17:07:18 Acked-by: Martin Jansa Oct 22 17:07:18 Acked-by: Chris Larson Oct 22 17:07:18 03Frans Meulenbroeks  07org.openembedded.dev * r218e3511f4 10openembedded.git/recipes/kaffe/kaffe-gtk_1.1.5.bb: Oct 22 17:07:20 kaffe-gtk_1.1.5.bb: remove useless EXTRA_OECONF += "" Oct 22 17:07:20 Signed-off-by: Frans Meulenbroeks Oct 22 17:07:20 Acked-by: Martin Jansa Oct 22 17:07:20 Acked-by: Chris Larson Oct 22 17:07:21 03Frans Meulenbroeks  07org.openembedded.dev * re18d963d8e 10openembedded.git/recipes/kaffe/kaffe-gtk_cvs.bb: Oct 22 17:07:21 kaffe-gtk_cvs.bb: remove useless EXTRA_OECONF += "" Oct 22 17:07:22 Signed-off-by: Frans Meulenbroeks Oct 22 17:07:33 Signed-off-by: Frans Meulenbroeks Oct 22 17:07:33 Acked-by: Martin Jansa Oct 22 17:07:34 Acked-by: Chris Larson Oct 22 17:10:28 okay, we need DISTRO_FEATURES for documentation stuff, anyone have any thoughts on this? Oct 22 17:10:44 Do we want one big docs feature, or separate man pages from other docs? Oct 22 17:10:48 i'm thinking just one for now Oct 22 17:16:40 * Crofton|work beats kergoth for missing OEDEM :) Oct 22 17:16:47 * kergoth is beaten Oct 22 17:17:01 I need to propose OADEM Oct 22 17:17:04 er Oct 22 17:17:06 OEDAM Oct 22 17:17:41 hopefully can make the one next year, this year was just bad timing Oct 22 17:18:28 yeah Oct 22 17:18:32 such is life Oct 22 17:18:50 I do think one on the left coast wuold be helpful though Oct 22 17:19:23 the developer base is getting further apart geo wise Oct 22 17:23:21 gah. anyone know what LD or gcc flag I'm missing when I get /lib/libc.so.6: file not recognized: File format not recognized ? Oct 22 17:23:30 I thought maybe -L or -B Oct 22 17:23:34 but those look ok Oct 22 17:25:37 nm I think I got it Oct 22 17:26:23 Howdy! Who's maintaining EFL's packages in OE? Oct 22 17:27:50 jconnolly: it doesnt' sound like something is missing, it sounds like somehting is there that shouldn't be (e.g. -L/usr/lib) Oct 22 17:29:58 thanks kergoth Oct 22 17:32:53 baah this merging is ugly Oct 22 17:33:11 kergoth: do you care about a clean commit history for tslib? Oct 22 17:34:04 morphis: clean history is nice in theory, but retaining history is nice as well, which is why i merged from you rather than rebasing the commits on master, and of course, there's practical concerns Oct 22 17:34:17 ok Oct 22 17:34:21 if its easier for you to merge from upstream master into your master and make that for-upstream, go for it Oct 22 17:35:33 it's easier but not the beautiful thing I want to have .. :) Oct 22 17:36:03 I have something ready now (http://github.com/playya/tslib/commits/for-upstream), but will test this first Oct 22 17:36:14 if it compiles fine .. Oct 22 17:36:47 k Oct 22 17:37:41 i like rebasing stuff thats for upstream, but as an upstream, i like merging it in -- i think that might be a decent rule of thumb Oct 22 17:37:46 * kergoth ponders Oct 22 17:38:05 jepp I even like rebasing stuff for upstream Oct 22 17:38:53 but on some point in time our tslib repository messy ... maybe because some more people worked on it Oct 22 17:38:54 but in this case an exception is appropriate -- oftentimes rebasing causes more headaches due to the repeated conflicts Oct 22 17:39:00 * kergoth nods Oct 22 17:39:03 maybe Oct 22 17:39:12 well, its had multiple conversions Oct 22 17:39:17 it went from cvs to svn, then from svn to git Oct 22 17:39:21 that probably doesnt help Oct 22 17:39:24 but it makes life easier Oct 22 17:39:28 and has gone through a few different maintainers Oct 22 17:39:32 * kergoth hugs git Oct 22 17:40:07 hehe ... jepp git is maybe the best FOSS invention of the last years Oct 22 17:41:34 mercurial seems nice too, but i haven't used it as much Oct 22 17:42:20 * kergoth should really sit down and fix X being pulled in by task-base Oct 22 17:43:20 03Jaap de Jong  07master * r93c4d9d840 10openembedded.git/recipes/cacao/ (3 files): Oct 22 17:43:20 cacao: Fix path for cacaoh source in do_install Oct 22 17:43:20 Signed-off-by: Jaap de Jong Oct 22 17:43:20 Signed-off-by: Khem Raj Oct 22 17:45:54 morphis: note that i don't even have a linux device handy with a touchscreen anymore, so am not in a position to do much tslib testing anymore -- if one of you folk want to take over the Maintainer role, I'm willing to hand it over :) Oct 22 17:46:31 kergoth: jepp I already heard that Oct 22 17:47:35 and thinking about your offer Oct 22 17:55:27 Crofton|work: yeah OEDAM all for it Oct 22 17:55:34 gm all Oct 22 17:55:41 gm Oct 22 17:55:53 Crofton|work: we can schedule one around ELC in 2011 Oct 22 17:56:08 or ESC in May Oct 22 17:56:09 we should think about that Oct 22 17:56:23 right. Oct 22 17:57:27 morphis: k. no big deal, i don't mind keeping an eye on it, applying patches people send isn't difficult :) Oct 22 17:57:34 Crofton|work: ELC will be a good shot Oct 22 17:57:53 kergoth: jepp Oct 22 17:58:15 kergoth: but I am not very familiar with the deeper internals of tslib Oct 22 17:58:18 kergoth: yes if you fix that X getting into everything, I can owe you some beers Oct 22 17:58:33 morphis: k, i don't mind keeping it Oct 22 17:58:51 khem: heh. i put some work into it while i was at MV, but i don't recall exactly what i did, and i don't think it got pushed, and now i'm not there, so .. :) Oct 22 17:59:04 and it seems nobody else there is pushing anything Oct 22 17:59:39 kergoth: btw. you can now pull again from playya/tslib:master or playya/tslib:for-upstream Oct 22 17:59:42 kergoth: hm Oct 22 18:00:18 k Oct 22 18:00:43 done, and pushed Oct 22 18:00:54 kergoth: but I thinking about taking maintainership for tslib Oct 22 18:00:57 great Oct 22 18:00:59 thank you Oct 22 18:01:08 k Oct 22 18:01:09 np Oct 22 18:06:09 kergoth: we need to sit and look into dot files Oct 22 18:07:11 yes Oct 22 18:07:13 03Simon Busch  07org.openembedded.dev * rc0155f9063 10openembedded.git/recipes/tslib/ (tslib/include-cy8mrln-header.patch tslib_git.bb): Oct 22 18:07:13 tslib: bump git SRCREV and update include-cy8mrln-header.patch accordingly Oct 22 18:07:13 Signed-off-by: Simon Busch Oct 22 18:50:17 http://wiki.openembedded.org/index.php/Darwin - starting to collect info about the setup Oct 22 19:04:35 gnumail says missing /scratch/oe/sysroots/x86_64-linux/usr/bin/java Oct 22 19:04:55 which which package should provide java Oct 22 19:23:41 kergoth: I get the following error when now compiling tslib in OE Oct 22 19:23:42 Initialized empty Git repository in /home/morphis/oe/sources/git/github.com.kergoth.tslib.git/.git/ Oct 22 19:23:42 remote: error: Could not read 25451aac6187e9d03461b5312642a74f3f617c6e Oct 22 19:23:42 remote: fatal: Failed to traverse parents of commit ea5afdd210d31fb6d8e3918694090070c764b978 Oct 22 19:23:42 remote: aborting due to possible repository corruption on the remote side. Oct 22 19:24:05 any idea? Oct 22 19:24:57 maybe you have to push again? Oct 22 19:25:07 as there is something missing? Oct 22 19:29:05 maybe git push origin --all helps? Oct 22 19:36:05 03Martin Jansa  07org.openembedded.dev * r0770548c90 10openembedded.git/conf/bitbake.conf: (log message trimmed) Oct 22 19:36:05 bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty Oct 22 19:36:05 * empty MACHINE_OVERRIDES adds :: to OVERRIDES (then in FILESPATH skips Oct 22 19:36:05 every directory after this) Oct 22 19:36:05 * strip() is needed to remove space in case some distribution has ie Oct 22 19:36:06 MACHINE_OVERRIDES ?= "" Oct 22 19:36:07 and then Oct 22 19:38:58 morphis: er.. i have no idea how that could happen Oct 22 19:39:07 hmm Oct 22 19:39:11 pushing doesn't help Oct 22 19:39:35 what the hell happened Oct 22 19:40:26 khem: ping Oct 22 19:43:16 khem: Crofton|work and more commiters: I'm working with a libz160-bin an another repo's of oe-imx stuffs. I cloned tardyp's repo and add a lot of stuffs. Is it interesting to keep in official OE repo? Oct 22 20:35:15 fidencio: publish the patches for review and if they are good we will pick them Oct 22 20:38:53 khem: fyi, http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=6901dc958c4254bafd4105c6c751246d50ba5ecd changes CURL_FEATURES assignment to unconditional... Oct 22 20:39:46 it was conditional so it can be set from distro/local.conf... now it doesn't work that way Oct 22 20:40:25 denix: hmmm Oct 22 20:40:41 not even if you do CURL_FEATURES_local = "" Oct 22 20:41:08 we need virtclass-target override Oct 22 20:41:24 like other virtclass overrides Oct 22 20:41:38 then we can solve this problem easily and many other problems. Oct 22 20:42:07 sometimes you want to use BBCLASSEXTEND but target recipe needs more stuff in a variable then the native one Oct 22 20:42:21 there is no way to append more stuff in such cases Oct 22 20:42:32 vice versa works ok Oct 22 20:42:34 ah, forgot about _local override, let me try Oct 22 20:43:07 yeah, I understand the reason - we don't have _virtcalls-target override... Oct 22 20:43:35 _virtclass-target I mean :) Oct 22 20:43:48 whats the reason not to have it ? Oct 22 20:45:13 I understand the reason for your change, because we don't have it. I don't know why we don't have it - wasn't envisioned during BBCLASSEXTEND design/implementation? :) Oct 22 20:45:47 i see Oct 22 20:45:58 I thought you understood the reason not to have it Oct 22 20:52:16 virtclass-target wouldn't make any sense -- target isn't generated by bbclassextend Oct 22 20:52:22 we can add a target override easily enough, though Oct 22 20:53:00 kergoth: I was trying to be consistent naming does not matter Oct 22 20:53:17 but the ability to do it is more important Oct 22 20:53:25 kergoth: could you cook up something ? Oct 22 20:53:42 remote: error: Could not read 25451aac6187e9d03461b5312642a74f3f617c6e Oct 22 20:53:42 remote: fatal: Failed to traverse parents of commit ea5afdd210d31fb6d8e3918694090070c764b978 Oct 22 20:53:42 remote: aborting due to possible repository corruption on the remote side. Oct 22 20:54:04 nschle85: something is broken on the server end of the tslib git repository, i need to see if the github folks can fix it Oct 22 20:54:35 khem: OVERRIDES =. "${@['', 'target:']['${PN}' == '${BPN}']}" Oct 22 20:55:19 kergoth: cool may be we should add it to bitbake.conf ? Oct 22 20:55:34 if it works, i'd think that would be a good idea, yeah Oct 22 20:55:50 lemme try it with curl Oct 22 20:56:26 an alternative would be to make it truly generic based on the base name, strip off the bits after BPN in the PN, the postfix, and remove the leading -, and add that to overrides -- if its empty, set to target -- then native/sdk/etc would be in overrides whether you use bbclassextend or not Oct 22 20:58:19 hmm Oct 22 20:58:58 then we can drop those virtclass_* overrides Oct 22 20:59:11 but it could be slower Oct 22 20:59:29 as it has to be computed for every recipe now ? Oct 22 21:02:39 kergoth: we have base_prune_suffix Oct 22 21:03:18 i know, thats how BPN gets set Oct 22 21:16:17 we use the 'proto' url param for svn, but the 'protocol' one for git? Oct 22 21:16:18 * kergoth sighs Oct 22 21:19:09 kergoth: ok I added a new function to return the suffix that matches otherwise return target Oct 22 21:19:20 but now how to prepend to OVERRIDES is tricky Oct 22 21:19:35 kergoth: OVERRIDES =. barfs Oct 22 21:19:46 and OVERRIDES_prepend errs out too Oct 22 21:20:11 kergoth: is there some other place that gets to each recipe during parsing besides bitbake.conf ? Oct 22 21:21:36 khem: In my specific case, I'm using packages from freescale that needed of a "Accept and Understand that bla bla bla" ... Some problem of packages like this in the tree? Oct 22 21:24:18 kergoth: this seems to work http://pastebin.com/RCtL4Pm2 Oct 22 21:24:29 kergoth: I dont like that prepend though Oct 22 21:24:51 hello khem, kergoth Oct 22 21:24:58 ant_: hi Oct 22 21:25:07 CrazyFoam: ping Oct 22 21:25:10 fidencio: ? Oct 22 21:26:10 hi Marex Oct 22 21:26:43 khem: Some problems in keep in the tree freescale binaries? afaik, the binaries are free but I don't know if are opensource Oct 22 21:27:09 hmmm no binaries are frowned upon Oct 22 21:27:14 get the sources Oct 22 21:27:49 khem: as I was thinking ... Oct 22 21:28:12 you can keep them in your own repo no problems Oct 22 21:28:36 khem: yeap. I'm doing it :-) Oct 22 21:30:00 kergoth: if a package is provided by multiple recipes what does bitbake world do Oct 22 21:30:09 will it pick one at random Oct 22 21:38:49 khem: is a build from scratch supposed to finish atm ? Oct 22 21:39:24 pending issues? Oct 22 21:40:12 03Michael Smith  07master * r117024cefa 10openembedded.git/recipes/xdelta/xdelta_3.0z.bb: Oct 22 21:40:12 xdelta 3.0z: add Oct 22 21:40:12 Signed-off-by: Michael Smith Oct 22 21:40:12 Cc: Sergei Frankoff Oct 22 21:43:16 ant_: all should work fine Oct 22 21:43:29 ant_: I have build x11-image and native-sdk-image without any issues Oct 22 21:47:06 sounds great Oct 22 21:47:24 I'll compile some kernels and rebuild all then Oct 22 21:47:50 khem: outstanding work ! Oct 22 21:48:04 keep on ;) Oct 22 22:42:39 khem: no, bitbake builds the preferred recipe when something is provided by more than one, it knows not to build both Oct 22 22:42:55 of course, if no recipe is preferred, its behavior is undefined Oct 22 22:43:19 I'm having trouble with variable conditional appending - not sure what I'm doing wrong: I have an u-load_git.bbappend to add some patches to u-load for my omap3 SOC board, so I SRC_URI_append_omap3 += "foo" yet it doesn't get added to SRC_URI - the same thing 'does' work for x-load_git.bbappend however Oct 22 22:43:55 i'm not seeing the point in using += when you're already appending with _append Oct 22 22:43:59 redundant at best Oct 22 22:44:01 been trying to use bitbake -e to troubleshoot but not seeing the issue - note that x-load_git.bb does have a SRC_URI_omap3 yet u-boot_git.bb does not Oct 22 22:44:10 thats why. Oct 22 22:44:18 your'e appending to SRC_URI, which gets overwritten by SRC_URI_omap3 Oct 22 22:45:04 something not clear from the bitbake manual: SRC_URI_append_foo - wouldn't it 'overrite' it each time so if I'm appending to a recipe and don't know if they already have a SRC_URI_append_foo - I want to 'add' to it - thats why the += Oct 22 22:45:20 no, it doesnt "overwrite" anything Oct 22 22:45:26 SRC_URI_append_foo isn't the name of a variable. Oct 22 22:45:38 SRC_URI_append_foo appends what you provide to SRC_URI when foo is in overrides Oct 22 22:45:49 SRC_URI_append = "foo" SRC_URI_append = "bar" Oct 22 22:45:52 the end result is "foobar" Oct 22 22:45:54 not "bar" Oct 22 22:46:04 append/prepend is an operation, not part of a variable name Oct 22 22:46:45 ok, that makes sense then - perhaps that example in bitbake manual would help Oct 22 22:48:00 so if SRC_URI_omap3 does not exist, a SRC_URI_append_omap3 will make it exist right? Oct 22 22:48:05 no. Oct 22 22:48:11 you're getting the order backwards. Oct 22 22:48:22 SRC_URI_append_omap3 appends to SRC_URI if omap3 Oct 22 22:48:27 it doesn't append to SRC_URI_omap3 Oct 22 22:48:33 if you want that, use SRC_URI_omap3_append Oct 22 22:49:09 ahhh... Oct 22 22:49:24 here's the thing Oct 22 22:49:27 append/prepend are operations Oct 22 22:49:35 adding an override after them makes the operation conditional Oct 22 22:49:44 but, an override specific varaible definition, like SRC_URI_omap3 Oct 22 22:49:47 that *is* the name of the variable Oct 22 22:49:55 but at the end of the parsing, that variable *replaces* SRC_URI Oct 22 22:50:15 ok that makes sense now, but still looking at my situation: my bbappends has SRC_URI_append_omap3 which should append to SRC_URI as I am omap3 - yet it doesn't Oct 22 22:50:29 so if you use the wrong one of __ vs __, you'll lose your changes entirely Oct 22 22:50:35 no, you're not listening Oct 22 22:50:43 step 1: it appends to SRC_URI Oct 22 22:50:49 step 2: SRC_URI_omap3 replaces SRC_URI Oct 22 22:50:56 ta da, your changes are gone Oct 22 22:51:07 this is why prepend/append suck Oct 22 22:51:16 order of operations inside of bitbake is important Oct 22 22:51:21 it forces us to be aware of how it does things Oct 22 22:51:33 * kergoth wants to see those two go away entirely Oct 22 22:51:51 but SRC_URI is 'not' getting 'replaced' by what I'm trying to append... Oct 22 22:52:03 ? Oct 22 22:52:08 no, its being replaced by SRC_URI_omap3 Oct 22 22:52:12 not by SRC_URI_append_omap3 Oct 22 22:52:19 you're still not grasping what i explained before Oct 22 22:52:28 it appends to SRC_URI, *NOT* SRC_URI_omap3 Oct 22 22:52:43 bitbake -e u-boot | grep ^SRC_URI= should this not show me what SRC_URI ends up being? perhaps I'm not getting bitbake -e Oct 22 22:52:53 yes, it does Oct 22 22:53:09 i already explained why your changes aren't in SRC_URI in bitbake -e Oct 22 22:53:14 i don't know how else to describe it Oct 22 22:53:22 just use SRC_URI_omap3_append instead and be happy Oct 22 22:53:47 heh... ok well I didn't understand that my changes are not shown in SRC_URI from bitbake -e Oct 22 22:54:25 guess that makes sense... bitbake -e isn't magic it just shows you all variables that have been set - 'post' append/prepend operations right? Oct 22 22:54:31 for like the 4th time, if you append to SRC_URI when SRC_URI_ is defined, your appended changes will not show up Oct 22 22:54:34 yes. Oct 22 22:54:37 in other words the operations occur dring bbparse? Oct 22 22:54:44 at the end of the recipe parsing, yes Oct 22 22:57:25 I think my big failing is I should not be so interested in SRC_URI from bitbake -e but be more interested in the overrides - for some reason I thought bitbake -e should have already applied the overrides the SRC_URI Oct 22 22:57:37 it has.. Oct 22 22:57:54 i'm not seeing what the problem is. you said what you appended wasn't in the SRC_URI in the bitbake -e, and i explained why Oct 22 22:58:11 you're appending wrong given the recipe Oct 22 22:58:32 bitbake -e is doing exactly what it should, it shows you the variables after all processing, including overrides Oct 22 23:01:11 let me ask another way then - (sorry thanks for the time!) - what is the 'correct' way to append to a SRC_URI from within a .bbappend? seems like it completely depends on what your appending to which kind of defeats the purpose of an append IMHO Oct 22 23:03:31 no, it doesn't. if you don't know the recipe you're appending to, you're not going to be able to accomplish much of anything. Oct 22 23:03:42 the recipe could always be using an override which would blow away your changes, or any number of other things Oct 22 23:03:56 it could use anonymous python to order pizza, for all i know Oct 22 23:03:56 what your saying makes complete sense its just not working for me - just figured out why - I was using an SOC override and the original recipe had a MACHINE override that was taking precidence - so yes it depends on the recipe Oct 22 23:04:07 if you havent read the recipe, you cant do shit and expect it to stick Oct 22 23:04:32 right, it just opens the door for something upstream to add an override that takes precidence over what your append is trying to do Oct 22 23:05:25 appending to a recipe like u-boot can be very tricky as there are a ton of overrides in u-boot Oct 22 23:05:52 so there really doesn't appear to be an easy way to some general patches for a variety of machines Oct 22 23:07:55 of course you can, as long as you know the recipe you're appending to. Oct 22 23:08:01 kergoth, at any rate thanks for the explanation - it did explain the nature of prepend/append to me that I didn't get before - I did have the correct syntax but was just using 'omap3' override when I needed to be using 'overo' override Oct 22 23:08:06 no problem Oct 22 23:08:10 ah Oct 22 23:08:20 right, I'm just saying that I have to add differently depending on the situation Oct 22 23:08:45 there are a boatload of overrides in u-boot - some based on machine, other based on SOC Oct 22 23:08:49 well, the only reason i brought htat all up is its a common misunderstanding that exemplifies one of the biggest problems with the file format -- parts are imperative, parts are declarative, and you have to be aware of when things happen Oct 22 23:09:24 ya, thats good - I did completely misunderstand append/prepend, but once I fixed it and still got wrong results I thought I was confused Oct 22 23:09:35 Crofton: ping Oct 22 23:09:38 * kergoth nods Oct 22 23:09:45 well, glad you got it sorted, anyway Oct 22 23:09:51 * kergoth can't wait for a new file format Oct 22 23:10:17 I take that back, looks like they are all machien overrides in u-boot recipe - which takes precidence over SOC so I shouldn't have been trying to use SOC override Oct 22 23:10:35 ah Oct 22 23:10:52 yet x-load didnt' have the machine override so was behaving differently Oct 22 23:10:56 * kergoth shakes head and grumbles Oct 22 23:10:59 I see the light Oct 22 23:11:01 yeah, that would explain it Oct 22 23:16:56 kergoth: do we copy m4 macros from Oct 22 23:17:02 during autoreconf Oct 22 23:17:07 tharvey: that's intended behaviour, isn't it? SOC is more generic than MACHINE Oct 22 23:17:14 khem: i don't think we pass the argument to aclocal for that Oct 22 23:17:22 not sure though Oct 22 23:17:24 hrmm Oct 22 23:17:39 aclocal --copy --force Oct 22 23:17:41 should do it Oct 22 23:18:16 kergoth: do u take care of that in your autotools branch ? Oct 22 23:20:04 i'm not sure what you're referring to. I don't see the need to copy the m4 files locally, the generated aclocal.m4 is sufficient -- the only reason to copy them locally is for distribution, and we don't distribute our source trees in that way Oct 22 23:21:15 so if I delete aclocal.m4 will it regenerate using the macros from sysroot Oct 22 23:22:56 yes Oct 22 23:23:11 but that isnt necessary, aclocal.m4 is removed in autotools.bbclass Oct 22 23:23:18 to ensure it gets reconstructed Oct 22 23:24:16 http://pastebin.com/3CCwcLdr Oct 22 23:24:35 that's advice for upstream maintainers Oct 22 23:24:40 its of no use to us Oct 22 23:24:54 but my libtool is wrongly generated Oct 22 23:25:01 if I copy those files to m4/ Oct 22 23:25:03 then it works Oct 22 23:25:35 well, dunno what to tell you, it pulls the macros from sysroot, *but* if they already had teh libtool macros copied into their tree, it'll likely prefer the local ones Oct 22 23:25:40 if you remove those, then it'll rely on sysroot alone Oct 22 23:25:48 best way is just devshell and run aclocal --verbose Oct 22 23:25:50 denix, yes, completely intended behavior just didn't think about it as I didn't realize there was a machien override in the base recipe Oct 22 23:25:52 you can see exactly where it pulls everything from Oct 22 23:26:18 kergoth: should base/utils functions be available from distro configs? Oct 22 23:26:43 kergoth: m4/ is empty Oct 22 23:26:45 any def'd function in a global place (read: class in INHERIT, or a .inc pulled in by a .conf) will be available everywhere Oct 22 23:26:49 still it gets it wronf Oct 22 23:26:53 khem: like i said, do the verbose, see where its coming from Oct 22 23:26:56 grep out libtool.m4 Oct 22 23:27:06 there is no libtool.4 Oct 22 23:27:14 well, whatever Oct 22 23:27:17 it copies ltmain.sh from the sysroot Oct 22 23:27:19 grep out the files you care about Oct 22 23:27:40 kergoth: what if it's used immediately evaluated variable? Oct 22 23:27:56 denix: then it'll fail unless the class that defines it was parsed before the variable definition Oct 22 23:28:36 kergoth: basically, I need to filter out INHERIT variable in a distro/local configs - any pointers? :) Oct 22 23:29:02 denix: base_filter_out is just a wrapper around oe.utils.filter_out, iirc Oct 22 23:29:06 look at utils.bbclass. Oct 22 23:29:12 just call the oe.utils version directly Oct 22 23:29:22 tried it already Oct 22 23:29:29 says oe is undeclared Oct 22 23:29:59 more specifically: NameError: name 'oe' is not defined Oct 22 23:30:01 ah, its running before the anonymous python in base.bbclass that imports it. Oct 22 23:30:07 do this, create a .inc Oct 22 23:30:09 in the .inc: Oct 22 23:30:13 def filter_my_crap(d): Oct 22 23:30:15 import oe.utils Oct 22 23:30:21 return oe.utils.filter_out(....) Oct 22 23:30:30 FOO := "${@filter_my_crap(d)}" Oct 22 23:30:37 then pull that in from your .conf Oct 22 23:30:44 kergoth: autoreconf --verbose --install --force --exclude=autopoint -I/usr/share/aclocal-1.11 -I /usr/share/aclocal Oct 22 23:30:55 aha, thanks Oct 22 23:30:56 this is a native recipe but still it should point to native sysroot Oct 22 23:31:06 when it comes to includes isnt it Oct 22 23:31:21 khem: the paths in autotools.bbclass don't point to /usr, as far as i can see Oct 22 23:31:42 then why does run.do_configure has it Oct 22 23:31:45 hmm Oct 22 23:32:16 good question, verys trange Oct 22 23:33:00 its a bitbake virtual/java-native Oct 22 23:33:55 kergoth: bitbake -e is here http://pastebin.com/imBxtGtx Oct 22 23:34:16 export prefix="/usr" Oct 22 23:34:56 that'd be the problem, natives shouldn't have a prefix set like that, because natives override STAGING_DIR_HOST to "" Oct 22 23:37:36 kergoth: ImportError: No module named oe.utils Oct 22 23:37:42 hmmm Oct 22 23:37:44 damnit Oct 22 23:38:00 all i can think of now is an anonymous python fuction Oct 22 23:38:06 get the var, adjust, set Oct 22 23:38:27 ok, thanks Oct 22 23:39:14 any ideas why oe.utils cannot be imported, though? Oct 22 23:40:01 yep, its in openembedded/lib Oct 22 23:40:08 whichisn't in sys.path until the base.bbclass bits adjust it Oct 22 23:40:11 basd on BBPATH Oct 22 23:40:14 blah, can't type Oct 22 23:40:45 got it Oct 22 23:41:05 anyway, filter_out is a one-liner anyway :) Oct 22 23:41:09 how much HDD space allocate to build container? :) Oct 22 23:41:26 :) Oct 22 23:41:31 Jay7: depends Oct 22 23:41:57 e.g. to build x11-image for all zauruses Oct 22 23:42:12 I hope 32Gb should be enough :) Oct 22 23:42:37 khem: Can I sent my patches for which ML? Oct 22 23:43:50 anyway this is lvm so I can resize later Oct 22 23:52:07 fidencio: oe Oct 22 23:52:10 look in woki Oct 22 23:52:12 wiki Oct 22 23:52:29 kergoth: why native.bbclass sets STAGING_DIR_HOST = "" Oct 22 23:56:36 yes, because prefix is already in staging Oct 22 23:56:40 if it didnt, you'd ahve the path twice Oct 23 00:02:45 khem: Can I send inline patches using format-patch? Oct 23 00:03:17 git send-mail Oct 23 00:03:19 err Oct 23 00:06:14 kergoth: this all mess is caused by the OVERRIDE_append :( Oct 23 00:06:46 somehow the var CONFIGUREAPPEND gets expanded before native is inherited Oct 23 00:10:13 so we need to move it to some place else Oct 23 01:44:17 kergoth: any hints Oct 23 01:44:26 for the OVERRIDES_append **** ENDING LOGGING AT Sat Oct 23 02:59:57 2010